Classical Machine Learning Amazon.png
Jeff Wilke, Amazon's consumer worldwide CEO, delivering a keynote presentation at re:MARS 2019

The history of Amazon's recommendation algorithm

Collaborative filtering and beyond.

In 2017, when the journal IEEE Internet Computing was celebrating its 20th anniversary, its editorial board decided to identify the single paper from its publication history that had best withstood the “test of time”. The honor went to a 2003 paper called “Amazon.com Recommendations: Item-to-Item Collaborative Filtering”, by then Amazon researchers Greg Linden, Brent Smith, and Jeremy York.

Collaborative filtering is the most common way to do product recommendation online. It’s “collaborative” because it predicts a given customer’s preferences on the basis of other customers’.

“There was already a lot of interest and work in it,” says Smith, now the leader of Amazon’s Weblab, which does A/B testing (structured testing of variant offerings) at scale to enable data-driven business decisions. “The world was focused on user-based collaborative filtering. A user comes to the website: What other users are like them? We sort of turned it on its head and found a different way of doing it that had a lot better scaling and quality characteristics for online recommendations.”

The better way was to base product recommendations not on similarities between customers but on correlations between products. With user-based collaborative filtering, a visitor to Amazon.com would be matched with other customers who had similar purchase histories, and those purchase histories would suggest recommendations for the visitor.

With item-to-item collaborative filtering, on the other hand, the recommendation algorithm would review the visitor’s recent purchase history and, for each purchase, pull up a list of related items. Items that showed up repeatedly across all the lists were candidates for recommendation to the visitor. But those candidates were given greater or lesser weight depending on how related they were to the visitor's prior purchases.

That notion of relatedness is still derived from customers’ purchase histories: item B is related to item A if customers who buy A are unusually likely to buy B as well. But Amazon’s Personalization team found, empirically, that analyzing purchase histories at the item level yielded better recommendations than analyzing them at the customer level.

Family ties

Beyond improving recommendations, item-to-item collaborative filtering also offered significant computational advantages. Finding the group of customers whose purchase histories most closely resemble a given visitor’s would require comparing purchase histories across Amazon’s entire customer database. That would be prohibitively time consuming during a single site visit.

The history of Amazon's recommendation algorithm | Amazon Science

The alternatives are either to randomly sample other customers in real time and settle for the best matches found or to build a huge offline similarity index by comparing every customer to every other. Because Amazon customers’ purchase histories can change dramatically in the course of a single day, that index would have to be updated regularly. Even offline indexing presents a huge computational burden.

On average, however, a given product sold on Amazon.com is purchased by only a tiny subset of the site’s customers. That means that inspecting the recent-purchase histories of everyone who bought a given item requires far fewer lookups than identifying the customers who most resemble a given site visitor. Smith and his colleagues found that even with early-2000s technology, it was computationally feasible to produce an updated list of related items for every product on the Amazon site on a daily basis.

The crucial question: how to measure relatedness. Simply counting how often purchasers of item A also bought item B wouldn’t do; that would make a few bestsellers like Harry Potter books and trash bags the top recommendations for every customer on every purchase.

Instead, the Amazon researchers used a relatedness metric based on differential probabilities: item B is related to item A if purchasers of A are more likely to buy B than the average Amazon customer is. The greater the difference in probability, the greater the items’ relatedness.

When Linden, Smith, and York published their paper in IEEE Internet Computing, their item-based recommendation algorithm had already been in use for six years. But it took several more years to identify and correct a fundamental flaw in the relatedness measure.

Getting the math right

The problem: the algorithm was systematically underestimating the baseline likelihood that someone who bought A would also buy B. Since a customer who buys a lot of products is more likely to buy A than a customer who buys few products, A buyers are, on average, heavier buyers than the typical Amazon customer. But because they’re heavy buyers, they’re also unusually likely to buy B.

Smith and his colleagues realized that it wasn’t enough to assess the increased likelihood of buying product B given the purchase of product A; they had to assess the increased likelihood of buying product B with any given purchase. That is, they discounted heavy buyers’ increased likelihood of buying B according to the heaviness of their buying.

“That was a large improvement to recommendations quality, when we got the math right,” Smith says.

That was more than a decade ago. Since then, Amazon researchers have been investigating a wide variety of ways to make customer recommendations more useful: moving beyond collaborative filtering to factor in personal preferences such as brands or fashion styles; learning to time recommendations (you may want to order more diapers!); and learning to target recommendations to different users of the same account, among many other things.

In June 2019, during a keynote address at Amazon’s first re:MARS conference, Jeff Wilke, the CEO of Amazon’s consumer division, highlighted one particular advance, in the algorithm for recommending movies to Amazon’s Prime Video customers. Amazon researchers’ innovations led to a twofold improvement in that algorithm’s performance, which Wilke described as a “once-in-a-decade leap”.

Entering the matrix

Recommendation is often modeled as a matrix completion problem. Imagine a huge grid, whose rows represent Prime Video customers and whose columns represent the movies in the Prime Video catalogue. If a customer has seen a particular movie, the corresponding cell in the grid contains a one; if not, it’s blank. The goal of matrix completion is to fill in the grid with the probabilities that any given customer will watch any given movie.

In 2014, Vijai Mohan’s team in the Personalization group — Avishkar Misra, Jane You, Rejith Joseph, Scott Le Grand, and Eric Nalisnick — was asked to design a new recommendation algorithm for Prime Video. At the time, the standard technique for generating personalized recommendations was matrix factorization, which identifies relatively small matrices that, multiplied together, will approximate a much larger matrix.

Inspired by work done by Ruslan Salakhutdinov — then an assistant professor of computer science at the University of Toronto — Mohan’s team instead decided to apply deep neural networks to the problem of matrix completion.

The typical deep neural network contains thousands or even millions of simple processing nodes, arranged into layers. Data is fed into the nodes of the bottom layer, which process it and pass their results to the next layer, and so on; the output of the top layer represents the result of some computation.

Training the network consists of feeding it lots of sample inputs and outputs. During training, the network’s settings are constantly adjusted, until they minimize the average discrepancy between the top layer’s output and the target outputs in the training examples.

Reconstruction

Matrix completion methods commonly use a type of neural network called an autoencoder. The autoencoder is trained simply to output the same data it takes as input. But in-between the input and output layers is a bottleneck, a layer with relatively few nodes — in this case, only 100, versus tens of thousands of input and output nodes.

As a consequence, the network can’t just copy inputs directly to outputs; it must learn a general procedure for compressing and then re-expanding every example in the training set. The re-expansion will be imperfect: in the movie recommendation setting, the network will guess that customers have seen movies they haven’t. But when, for a given customer-movie pair, it guesses wrong with high confidence, that’s a good sign that the customer would be interested in that movie.

To benchmark the autoencoder’s performance, the researchers compared it to two baseline systems. One was the latest version of Smith and his colleagues’ collaborative-filtering algorithm. The other was a simple listing of the most popular movie rentals of the previous two weeks. “In the recommendations world, there’s a cardinal rule,” Mohan says. “If I know nothing about you, then the best things to recommend to you are the most popular things in the world.”

To their mild surprise, the item-to-item collaborative-filtering algorithm outperformed the autoencoder. But to their much greater surprise, so did the simple bestseller list. The autoencoder’s performance was “so bad that we had to go and doublecheck and re-run the experiments multiple times,” Mohan says. “I was giving a hard time to the scientists. I was saying, ‘You probably made a mistake.’”

Once they were sure the results were valid, however, they were quick to see why. In a vacuum, matrix completion may give the best overview of a particular customer’s tastes. But at any given time, most movie watchers will probably opt for recent releases over neglected classics in their preferred genres.

Neural network classifiers with time considerations
Amazon researchers found that using neural networks to generate movie recommendations worked much better when they sorted the input data chronologically and used it to predict future movie preferences over a short (one- to two-week) period.
Amazon

So Mohan’s team re-framed the problem. They still used an autoencoder, but they trained it on movie-viewing data that had been sorted chronologically. During training, the autoencoder saw data on movies that customers had watched before some cutoff time. But it was evaluated on how well it predicted the movies they had watched in the two-week period after the cutoff time.

Because Prime Video’s Web interface displays six movie recommendations on the page associated with each title in its catalogue, the researchers evaluated their system on whether at least one of its top six recommendations for a given customer was in fact a movie that that customer watched in the two-week period after the cutoff date. By that measure, not only did the autoencoder outperform the bestseller list, but it also outperformed item-to-item collaborative filtering, two to one. As Wilke put it at re:MARS, “We had a winner.”

Whether any of the work that Amazon researchers are doing now will win test-of-time awards two decades hence remains to be seen. But Smith, Mohan, and their colleagues will continue to pursue new approaches to designing recommendation algorithms, in the hope of making Amazon.com that much more useful for customers.


Amazon Science Newsletter Project Kuiper.jpg
Get more from Amazon Science
Sign up for our monthly newsletter

Work with us

See More Jobs
MX, DIF, Mexico City
At Amazon Web Services (AWS), we’re hiring highly technical Data and Machine Learning engineers to collaborate with our customers and partners on key engagements. Our consultants will develop and deliver proof-of-concept projects, technical workshops, and support implementation projects. These professional services engagements will focus on customer solutions such as Machine Learning, Data and Analytics, HPC and more.In this role, you will work with our partners, customers and focus on our AWS offerings such Amazon Kinesis, AWS Glue, Amazon Redshift, Amazon EMR, Amazon Athena, Amazon SageMaker and more. You will help our customers and partners to remove the constraints that prevent them from leveraging their data to develop business insights.AWS Professional Services engage in a wide variety of projects for customers and partners, providing collective experience from across the AWS customer base and are obsessed about customer success. Our team collaborates across the entire AWS organization to bring access to product and service teams, to get the right solution delivered and drive feature innovation based upon customer needs.You will also have the opportunity to create white papers, writing blogs, build demos and other reusable collateral that can be used by our customers. Most importantly, you will work closely with our Solution Architects, Data Scientists and Service Engineering teams.The ideal candidate will have extensive experience with design, development and operations that leverages deep knowledge in the use of services like Amazon Kinesis, Apache Kafka, Apache Spark, Amazon SageMaker, Amazon EMR, NoSQL technologies and other 3rd parties.This is a customer facing role. You will be required to travel to client locations and deliver professional services when needed.
MX, DIF, Mexico City
At Amazon Web Services (AWS), we’re hiring highly technical Data and Machine Learning engineers to collaborate with our customers and partners on key engagements. Our consultants will develop and deliver proof-of-concept projects, technical workshops, and support implementation projects. These professional services engagements will focus on customer solutions such as Machine Learning, Data and Analytics, HPC and more.In this role, you will work with our partners, customers and focus on our AWS offerings such Amazon Kinesis, AWS Glue, Amazon Redshift, Amazon EMR, Amazon Athena, Amazon SageMaker and more. You will help our customers and partners to remove the constraints that prevent them from leveraging their data to develop business insights.AWS Professional Services engage in a wide variety of projects for customers and partners, providing collective experience from across the AWS customer base and are obsessed about customer success. Our team collaborates across the entire AWS organization to bring access to product and service teams, to get the right solution delivered and drive feature innovation based upon customer needs.You will also have the opportunity to create white papers, writing blogs, build demos and other reusable collateral that can be used by our customers. Most importantly, you will work closely with our Solution Architects, Data Scientists and Service Engineering teams.The ideal candidate will have extensive experience with design, development and operations that leverages deep knowledge in the use of services like Amazon Kinesis, Apache Kafka, Apache Spark, Amazon SageMaker, Amazon EMR, NoSQL technologies and other 3rd parties.This is a customer facing role. You will be required to travel to client locations and deliver professional services when needed.
US, CA, Sunnyvale
As a Senior Machine Learning engineer you will be responsible for developing, optimizing, testing, and deploying hardware edge machine learning solutions used in a range of Amazon devices. You will collaborate closely with a team of machine learning scientists and hardware designers to implement solutions in various areas including speech, vision, robotic perception and sensor fusion. You will also help drive the system architecture, spearhead best practices that enable a quality product, and help coach and develop junior engineers.In this role you will:· Analyze machine learning models for deployment in Amazon device hardware accelerators.· Train and optimize models to work most efficiently on Amazon devices· Develop processes and software automation to automate mapping of machine learning models onto Amazon device hardware accelerators· Work with hardware architects and engineers to implement hardware acceleration for Amazon devices
US, WA, Seattle
The Amazon Air Science and Technology team is seeking a Data Scientist to be part of a team solving complex aviation operations problems to reduce cost and improve performance. This is a blue-sky role that gives you a chance to bring optimization modeling, statistical modeling, machine learning advancements to data analytics for customer-facing solutions in complex industrial settings.You will work closely with product, research science and technical leaders throughout Amazon Air, Amazon Delivery Technology and Supply Chain Optimization and will be responsible for influencing funding decisions in areas of investment that you identify as critical future product offerings. You will partner with software developers and data scientists to build end-to-end data pipelines and production code, and you will have exposure to senior leadership as we communicate results and provide scientific guidance to the business. You will analyze large amounts of business data, build the machine learning or optimization models that will enable us to continually delight our customers worldwide.The ideal candidate will have extensive experience in Science work, business analytics and have the aptitude to incorporate new approaches and methodologies while dealing with ambiguities. Excellent business and communication skills are a must to develop and define key business questions and to build data sets that answer those questions. You should have a demonstrated ability to think strategically and analytically about business, product, and technical challenges. Further, you must have the ability to build and communicate compelling value propositions, and work across the organization to achieve consensus. This role requires a strong passion for customers, a high level of comfort navigating ambiguity, and a keen sense of ownership and drive to deliver results.RESPONSIBILITIES:· Utilize code (Python, R, Scala, etc.) for analyzing data and building statistical models to solve specific business problems.· Improve upon existing methodologies by developing new data sources, testing model enhancements, and fine-tuning model parameters.· Collaborate with researchers, software developers, and business leaders to define product requirements and provide analytical support· Directly contribute to the design and development of automated selection systems· Build customer-facing reporting tools to provide insights and metrics which track system performance· Communicate verbally and in writing to business customers and leadership team with various levels of technical knowledge, educating them about our systems, as well as sharing insights and recommendations
US, WA, Seattle
The Amazon Air Science and Technology team is seeking an Applied Scientist to be part of a team solving complex aviation operations problems to reduce cost and improve performance. This is a blue-sky role that gives you a chance to bring optimization modeling, statistical modeling, machine learning advancements to data analytics for customer-facing solutions in complex industrial settings.You will work closely with product, research science and technical leaders throughout Amazon Air, Amazon Delivery Technology and Supply Chain Optimization and will be responsible for influencing funding decisions in areas of investment that you identify as critical future product offerings. You will partner with software developers and data scientists to build end-to-end data pipelines and production code, and you will have exposure to senior leadership as we communicate results and provide scientific guidance to the business. You will analyze large amounts of business data, build the machine learning or optimization models that will enable us to continually delight our customers worldwide.The ideal candidate will have extensive experience in Science work, business analytics and have the aptitude to incorporate new approaches and methodologies while dealing with ambiguities. Excellent business and communication skills are a must to develop and define key business questions and build models that answer those questions. You should have a demonstrated ability to think strategically and analytically about business, product, and technical challenges. Further, you must have the ability to build and communicate compelling value propositions, and work across the organization to achieve consensus. This role requires a strong passion for customers, a high level of comfort navigating ambiguity, and a keen sense of ownership and drive to deliver results.Tasks/ Responsibilities:· Partnership with the engineering and operations to drive modeling and design for complex business problems.· Design and prototype decision support tools (product) to automate standardized processes and optimize trade-offs across the full decision space.· Contribute to the mid- and long-term strategic planning studies and analysis.· Lead complex transportation modeling analyses to aid management in making key business decisions and set new policies.
US, WA, Seattle
Want to watch a movie at the end of a long week, but not sure what to choose? Looking for a new show while you wait for the next season of Game of Thrones to start? So are millions of our Prime Video customers. The Prime Video Relevance team helps customers find relevant videos, channels and topics so they can find content they didn’t even known they were looking for, continuing to surprise them with the depth of our catalog.We tailor our recommendations through a variety of machine learning algorithms including deep learning neural networks, that you will help define and extend. We are looking for creative, customer and details obsessed machine learning scientists who can apply the latest research, state of the art algorithms and machine learning to build highly scalable recommendation and personalization systems. You'll have a chance to collaborate with talented teams of engineers and scientists to run these predictions on distributed systems at incredible scale and speed.As a member of the Prime Video Personalization organization, you will spend your time as a hands-on machine learning practitioner and a research leader. You will play a key role on the team, building and guiding machine learning models from the ground up. At then of the day, you will have the reward of seeing your contributions benefit millions of Amazon.com customers worldwide.Some examples of the things we work on:· Using Neural Networks and Deep Learning techniques to find titles that customers will enjoy· Build and operate services that deliver millions of recommendations per second· Extend models and algorithms to support our ever growing ways of consuming content (subscriptions, live, rentals etc), dealing with unique challenges such as observational bias and rapidly scaling dimensions· Constantly experimenting with changes to the underlying algorithms and models to deliver relevant content to a wide variety of customer experiencesIf you are ready to truly make an impact on a product that is used by millions of people around the world, including your own friends and family, then we would love to talk to you.Amazon.com is an Equal Opportunity-Affirmative Action Employer – Minority / Female / Disability / Veteran / Gender Identity / Sexual Orientation
ES, B, Barcelona
If you get excited by the prospect of solving hard problems using Computer Vision and Machine Learning, enjoy working in a fast-paced environment with thought leaders in the CV space, and are passionate about launching algorithms for maximum customer impact, then we have the perfect role for you!We are looking for Principal Scientist with a deep expertise in computer vision to focus on data-driven image synthesis using deep learning techniques such as GANs and VAEs to help Amazon transform pixels into personalized fashion images. Our team is developing cutting-edge technology to personalize and transform photos, partnering with many different teams across Amazon to apply a mix of workflows, image generation, computer vision, and machine learning to solve problems like virtual try-on at Amazon scale. A strong candidate will have experience leading a team of scientists, working independently on problems, and setting their own research direction.As a Principal Applied Scientist, you will work in a team with other scientists and engineers working on products and prototypes in the field of image synthesis and photo-realistic appearance of people and clothing to create scalable solutions for business problems. You will play a critical role in ideation for the team and run live experiments, with opportunities to publish your work. We are building the next generation of fashion imagery, and we hope you'll join us!This role is located in Barcelona, Spain, but we are open to hiring remote positions across the US and EU.Amazon is committed to a diverse and inclusive workplace. Amazon is an equal opportunity employer and does not discriminate on the basis of race, national origin, gender, gender identity, sexual orientation, protected veteran status, disability, age, or other legally protected status. For individuals with disabilities who would like to request an accommodation, please visit https://www.amazon.jobs/en/disability/us.
US, WA, Seattle
The Human Resources Central Science Team (HRCS) uses economics, behavioral science, statistics, and machine learning to proactively identify mechanisms and process improvements which simultaneously improve Amazon and the lives, wellbeing, and the value of work to Amazonians. We are an interdisciplinary team that combines the talents of science and engineering to develop and deliver solutions that measurably achieve this goal.We are looking for economists who are able to provide structure around complex business problems, hone those complex problems into specific, scientific questions, and test those questions to generate insights. The ideal candidate will work with engineers and computer scientists to estimate models and algorithms on large scale data, design pilots and measure their impact, and transform successful prototypes into improved policies and programs at scale. We are looking for creative thinkers who can combine a strong technical economic toolbox with a desire to learn from other disciplines, and who know how to execute and deliver on big ideas as part of an interdisciplinary technical team.Ideal candidates will work closely with business partners to develop science that solves the most important business challenges. They will work in a team setting with individuals from diverse disciplines and backgrounds. They will serve as an ambassador for science and a scientific resource for business teams, so that scientific processes permeate throughout the HR organization to the benefit of Amazonians and Amazon. Ideal candidates will own the development of scientific models and manage the data analysis, modeling, and experimentation that is necessary for estimating and validating models. They will work closely with engineering teams to develop scalable data resources to support rapid insights, and take successful models and findings into production as new products and services. They will be customer-centric and will communicate scientific approaches and findings to business leaders, listening to and incorporate their feedback, and delivering successful scientific solutions.
US, WA, Seattle
The Human Resources Central Science Team (HRCS) uses economics, behavioral science, statistics, and machine learning to proactively identify mechanisms and process improvements which simultaneously improve Amazon and the lives, wellbeing, and the value of work to Amazonians. We are an interdisciplinary team that combines the talents of science and engineering to develop and deliver solutions that measurably achieve this goal.We are looking for economists who are able to provide structure around complex business problems, hone those complex problems into specific, scientific questions, and test those questions to generate insights. The ideal candidate will work with engineers and computer scientists to estimate models and algorithms on large scale data, design pilots and measure their impact, and transform successful prototypes into improved policies and programs at scale. We are looking for creative thinkers who can combine a strong technical economic toolbox with a desire to learn from other disciplines, and who know how to execute and deliver on big ideas as part of an interdisciplinary technical team.Ideal candidates will work closely with business partners to develop science that solves the most important business challenges. They will work in a team setting with individuals from diverse disciplines and backgrounds. They will serve as an ambassador for science and a scientific resource for business teams, so that scientific processes permeate throughout the HR organization to the benefit of Amazonians and Amazon. Ideal candidates will own the development of scientific models and manage the data analysis, modeling, and experimentation that is necessary for estimating and validating models. They will work closely with engineering teams to develop scalable data resources to support rapid insights, and take successful models and findings into production as new products and services. They will be customer-centric and will communicate scientific approaches and findings to business leaders, listening to and incorporate their feedback, and delivering successful scientific solutions.
US, WA, Seattle
The Amazon Product Catalog group is hiring a Senior Applied Scientist to help us build the most authoritative product knowledge in existence. Our charter is to capture every relevant fact and relationship for any product on the planet. You will have an enormous opportunity to make a large impact on the design, architecture, and implementation of cutting edge products used every day, by people you know.As a Senior Applied Scientist, you will be responsible for designing, developing, and deploying large-scale data mining solutions and distributed machine learning systems that ultimately make shopping on Amazon delightful and functional. Because the product catalog is the heart of our retail business, your work will directly change how Amazon customers search, find, compare, and buy everything from televisions to groceries. Our catalog contains hundreds of billions of facts, a scale which demands automated, state-of-the-art techniques to identify defects, abuse, incorrect data, and find new relationships between products, facts, and entities.Our domain blurs the boundaries between knowledge graphs, ontologies, entity recognition, image classification, machine translation, entity recognition, and semantic fact extraction. This requires a fast, collaborative environment We not only leverage best-in-class models, but also improve them. You will have the support of a well-established team with a long-term charter, science-smart engineers, language experts, and the latest AWS products and compute resources. You will collaborate closely with teams of software engineers, applied machine learning scientists, product managers, user interface designers, and others in order to influence our business and technical strategy, and play a key role in defining the team’s roadmap.A successful candidate will have an established background in machine learning science, large scale software systems, a strong technical ability, great communication skills, and a motivation to solve new, ambiguous problems.Unique Opportunities· Global impact on hundreds of millions of Amazon customers· Opportunities to publish both internally and externally· Close-knit partnership with science-smart developers – your improvements can roll out in days, not months, because engineers understand your work· Access to Amazon-scale datasets, spanning hundreds of billions of facts· Support from over 100 dedicated language experts and SMEs to label, validate, and test your hypothesis· Challenging, cutting-edge science problems that cross multiple domains (such as textual, semantic and image-aware Transformer models )Key Responsibilities· Develop production-ready machine learning solutions to improve Amazon Product Catalog· Collaborate with and influence scientists and engineers in multiple teams· Refine existing techniques, optimizing accuracy, throughput, and global impact· Publish results internally and externally
US, WA, Seattle
Interested in modeling and understanding customer behavior through machine learning, artificial intelligence, and data mining over TB scale data with huge business impact on millions of customers? Join our team of Scientists and Engineers developing models to predict customer behavior and optimize the customer experience with Amazon Prime. This includes identifying who our customers are and providing them with personalized relevant content. As an ML expert, you will partner directly with product owners to intake, build, and directly apply your modeling solutions.There are numerous scientific and technical challenges you will get to tackle in this role, such as global scalability of models, combinatorial optimization, cold start problem, accelerated experimentation, short/long term goals modeling, cohort identification in a semi-supervised setting, and multi-step optimization leading to reinforcement learning of the customer journey. We employ techniques from supervised learning, bandits, optimization, and RL.As the central science team within Prime, our expertise gets routinely called upon to weigh in on a variety of topics. We also emphasize the need and value of scientific research and have developed a strong publication and patent record (internally/externally) which you will be a part of.You will also utilize and be exposed to the latest in ML technologies and infrastructure: AWS technologies (EMR/Spark, Redshift, Sagemaker, DynamoDB, S3, ...), various ML algorithms and techniques (XGBoost, Random Forests, Neural Networks, supervised/unsupervised/semi-supervised/reinforcement learning), and statistical modeling techniques.Major responsibilities· · Build and develop machine learning models and supporting infrastructure at TB scale, in coordination with software engineering teams.· · Leverage Bandits and Reinforcement Learning for Recommendation Systems.· · Develop offline policy estimation tools and integrate with reporting systems.· · Establish scalable, efficient, automated processes for large scale data analyses, model development, model validation and model implementation.· · Analyze and extract relevant information from large amounts of Amazon’s historical business data to help automate and optimize key processes.· · Work closely with the business to understand their problem space, identify the opportunities and formulate the problems.· · Use machine learning, data mining, statistical techniques and others to create actionable, meaningful, and scalable solutions for the business problems.· · Design, develop and evaluate highly innovative models and statistical approaches to understand and predict customer behavior and to solve business problems.
US, WA, Seattle
Interested in modeling and understanding customer behavior through machine learning, artificial intelligence, and data mining over TB scale data with huge business impact on millions of customers? Join our team of Scientists and Engineers developing models to predict customer behavior and optimize the customer experience with Amazon Prime. This includes identifying who our customers are and providing them with personalized relevant content. As an ML expert, you will partner directly with product owners to intake, build, and directly apply your modeling solutions.There are numerous scientific and technical challenges you will get to tackle in this role, such as global scalability of models, combinatorial optimization, cold start problem, accelerated experimentation, short/long term goals modeling, cohort identification in a semi-supervised setting, and multi-step optimization leading to reinforcement learning of the customer journey. We employ techniques from supervised learning, bandits, optimization, and RL.As the central science team within Prime, our expertise gets routinely called upon to weigh in on a variety of topics. We also emphasize the need and value of scientific research and have developed a strong publication and patent record (internally/externally) which you will be a part of.You will also utilize and be exposed to the latest in ML technologies and infrastructure: AWS technologies (EMR/Spark, Redshift, Sagemaker, DynamoDB, S3, ...), various ML algorithms and techniques (XGBoost, Random Forests, Neural Networks, supervised/unsupervised/semi-supervised/reinforcement learning), and statistical modeling techniques.Major responsibilities· · Build and develop machine learning models and supporting infrastructure at TB scale, in coordination with software engineering teams.· · Leverage Bandits and Reinforcement Learning for Recommendation Systems.· · Develop offline policy estimation tools and integrate with reporting systems.· · Establish scalable, efficient, automated processes for large scale data analyses, model development, model validation and model implementation.· · Analyze and extract relevant information from large amounts of Amazon’s historical business data to help automate and optimize key processes.· · Work closely with the business to understand their problem space, identify the opportunities and formulate the problems.· · Use machine learning, data mining, statistical techniques and others to create actionable, meaningful, and scalable solutions for the business problems.· · Design, develop and evaluate highly innovative models and statistical approaches to understand and predict customer behavior and to solve business problems.
US, WA, Seattle
Global Talent Management (GTM) at Amazon owns a suite of products which helps drive career development for hundreds of thousands of Amazonians across the world. GTM - Science utilizes a wide array of data sources to conduct analytics and create predictive models that fuel recommendations, actions, and insights in nearly a dozen software systems. The team itself is composed of a variety of scientists and engineers with varied backgrounds, coming together to create diverse and innovative solutions to the problems faced by the one of the world’s largest and fastest growing workforces.This role will support the advancement of key workforce planning products owned by the team. The role will be a scientific lead for forecasting in the organization and a thought leader for forecasting applications throughout HR. If you’re interested in building models used regularly by thousands of Amazonians, to inform talent management decisions, this role is for you. You will support interesting, analytical problems, in an environment where you get to learn from other experienced economists and apply econometrics at massive scale.You will build econometric models, using our world class data systems, and apply economic theory to solve business problems in a fast moving environment. Economists at Amazon will be expected to develop new techniques to process large data sets, address quantitative problems, and contribute to design of automated systems around the company.· Build and operationalize econometric and statistical models· Perform model refreshes or updates to analyses as needed· Work collaboratively with economists and research scientists to assist in the design and implementation of analysis to answer challenging HR questions· Interpret and communicate results to outside customers· Aggregate and analyze data pulled from disparate sources (HR, Finance or other business systems) and related industry and external benchmarks; provide insights and a point of view on analysis and recommendations· Assist in the design and delivery of automated, scalable analytical models to stakeholders· Report results in a manner which is both statistically rigorous and compellingly relevant
US, VA, Arlington
Global Talent Management (GTM) at Amazon owns a suite of products which helps drive career development for hundreds of thousands of Amazonians across the world. GTM - Science utilizes a wide array of data sources to conduct analytics and create predictive models that fuel recommendations, actions, and insights in nearly a dozen software systems. The team itself is composed of a variety of scientists and engineers with varied backgrounds, coming together to create diverse and innovative solutions to the problems faced by the one of the world’s largest and fastest growing workforces.This role will support the advancement of key workforce planning products owned by the team. The role will be a scientific lead for forecasting in the organization and a thought leader for forecasting applications throughout HR. If you’re interested in building models used regularly by thousands of Amazonians, to inform talent management decisions, this role is for you. These are exciting fast-paced businesses in which work on extremely interesting analytical problems, in an environment where you get to learn from other experienced economists and apply econometrics at massive scale.You will build econometric models, using our world class data systems, and apply economic theory to solve business problems in a fast moving environment. Economists at Amazon will be expected to develop new techniques to process large data sets, address quantitative problems, and contribute to design of automated systems around the company.· Build and operationalize econometric and statistical models· Perform model refreshes or updates to analyses as needed· Work collaboratively with economists and research scientists to assist in the design and implementation of analysis to answer challenging HR questions· Interpret and communicate results to outside customers· Aggregate and analyze data pulled from disparate sources (HR, Finance or other business systems) and related industry and external benchmarks; provide insights and a point of view on analysis and recommendations· Assist in the design and delivery of automated, scalable analytical models to stakeholders· Report results in a manner which is both statistically rigorous and compellingly relevant
US, NY, New York
Machine learning (ML) has been strategic to Amazon from the early years. We are pioneers in areas such as recommendation engines, product search, eCommerce fraud detection, and large-scale optimization of fulfillment center operations.The Amazon ML Solutions Lab team helps AWS customers accelerate the use of machine learning to solve business and operational challenges and promote innovation in their organization. In this role, you will be designing and developing advanced ML models to solve diverse challenges and opportunities. You will be working with terabytes of text, images, and other types of data to solve real-world problems. You'll design and run experiments, research new algorithms, and find new ways of optimizing risk, profitability, and customer experience.We’re looking for talented data scientists capable of applying classical ML algorithms and cutting-edge deep learning (DL) and reinforcement learning approaches to areas such as drug discovery, customer segmentation, fraud prevention, capacity planning, predictive maintenance, pricing optimization, call center analytics, player pose estimation, event detection, and virtual assistant among others.The primary responsibilities of this role are to:· Design, develop, and evaluate innovative ML/DL models to solve diverse challenges and opportunities across industries· Interact with customer directly to understand their business problems, and help them with defining and implementing scalable ML/DL solutions to solve them· Work closely with account teams, research scientist teams, and product engineering teams to drive model implementations and new algorithmsThis position requires travel of up to 25%.Here at AWS, we embrace our differences. We are committed to furthering our culture of inclusion. We have ten employee-led affinity groups, reaching 40,000 employees in over 190 chapters globally. We have innovative benefit offerings, and we host annual and ongoing learning experiences, including our Conversations on Race and Ethnicity (CORE) and AmazeCon (gender diversity) conferences. Amazon’s culture of inclusion is reinforced within our 14 Leadership Principles, which remind team members to seek diverse perspectives, learn and be curious, and earn trust.
US, WA, Seattle
Are you excited about powering Amazon’s physical stores’ expansion through the application of Machine Learning and Big Data technologies? Do you thrive in a fast-moving, innovative environment that values data-driven decision making, scalable solutions, and sound scientific practices? We are looking for experienced scientists to build the next level of intelligence that will help Amazon physical stores grow and succeed.Our team is responsible for building the core intelligence, insights, and algorithms that support the real estate acquisition strategies for Amazon physical stores. We are tackling cutting-edge, complex problems — such as predicting the optimal location for new Amazon stores — by bringing together numerous data assets from disparate sources inside and outside of Amazon, and using best-in-class modeling solutions to extract the most information out of them.You will have a proven track-record of delivering solutions using advanced science approaches. You will be comfortable using a variety of tools and data sources to answer high-impact business questions. You will transform one-off models into automated systems. You will be able to break down complex information and insights into clear and concise language and be comfortable presenting your findings to audiences with a broad range of backgrounds.Responsibilities:· Develop production software systems utilizing advanced algorithms to solve business problems.· Analyze and validate data to ensure high data quality and reliable insights.· Partner with data engineering teams across multiple business lines to improve data assets, quality, metrics and insights.· Proactively identify interesting areas for deep dive investigations and future product development.· Design and execute experiments, and analyze experimental results in collaboration with Product Managers, Business Analysts, Economists, and other specialists.· Leverage industry best practices to establish repeatable applied science practices, principles & processes.
US, WA, Seattle
Amazon’s High Value Messaging (HVM) Analytics team (part of Customer Behavior Analytics) is looking for a Senior Applied Scientist to spearhead the rapid growth of our Marketing Measurement solutions. The team focuses on building scalable scientific models to estimate the effectiveness of Amazon marketing efforts and provide actionable insights to the various marketing teams within Amazon. We are looking for a thought leader that has an aptitude for delivering customer-focused solutions and who enjoys working on the intersection of Big-Data analytics, Machine/Deep Learning, and Causal Inference.A successful candidate will be a self-starter, comfortable with ambiguity, able to think big and be creative, while still paying careful attention to detail. You should be able to translate how data represents the customer journey, be comfortable dealing with large and complex data sets, and have experience using machine learning and econometric modeling to solve business problems. You should have strong analytical and communication skills, be able to work with product managers and software teams to define key business questions and work with the analytics team to solve them. You will join a highly collaborative and diverse working environment that will empower you to shape the future of Amazon marketing, as well as allow you to be part of the large science community within the Customer Behavior Analytics (CBA) organization.The Customer Behavior Analytics (CBA) organization owns Amazon’s insights pipeline, from data collection to deep analytics. We aspire to be the place where Amazon teams come for answers, a trusted source for data and insights that empower our systems and business leaders to make better decisions. Our outputs shape Amazon product and marketing teams’ decisions and thus how Amazon customers see, use, and value their experience.The main responsibilities for this position include:· Apply expertise in ML and causal modeling to develop systems that describe how Amazon’s marketing campaigns impact customers’ actions· Own the end-to-end development of novel scientific models that address the most pressing needs of our business stakeholders and help guide their future actions· Improve upon and simplify our existing solutions and frameworks· Review and audit modeling processes and results for other scientists, both junior and senior· Work with marketing leadership to align our measurement plan with business strategy· Formalize assumptions about how models are expected to behave, creating definitions of outliers, developing methods to systematically identify these outliers, and explaining why they are reasonable or identifying fixes for them· Identify new opportunities that are suggested by the data insights· Bring a department-wide perspective into decision making· Develop and document scientific research to be shared with the greater science community at Amazon
US, CA, San Diego
Do you want to join an innovative team of scientists who use machine learning and statistical techniques to help Amazon provide the best customer experience by protecting Amazon customers from hackers and bad actors? Do you want to build advanced algorithmic systems that help manage the trust and safety of millions of customer every day? Are you excited by the prospect of analyzing and modeling terabytes of data and create state-of-art algorithms to solve real world problems? Do you like to innovate and simplify? If yes, then you may be a great fit to join the Amazon Account Integrity team.The Amazon Account Integrity team works to ensure that customers are protected from bad actors trying to access their accounts. Our greatest challenge is protecting customer trust without unjustly harming good customers. To strike the right balance, we invest in mechanisms which allow us to accurately identify and mitigate risk, and to quickly correct and learn from our mistakes. This strategy includes continuously evolving enforcement policies, iterating our Machine Learning risk models, and exercising high‐judgement decision‐making where we cannot apply automation.
US, CA, San Diego
Do you want to join an innovative team of scientists who use machine learning and statistical techniques to help Amazon provide the best customer experience by protecting Amazon customers from hackers and bad actors? Do you want to build advanced algorithmic systems that help manage the trust and safety of millions of customer every day? Are you excited by the prospect of analyzing and modeling terabytes of data and create state-of-art algorithms to solve real world problems? Do you like to innovate and simplify? If yes, then you may be a great fit to join the Amazon Account Integrity team.The Amazon Account Integrity team works to ensure that customers are protected from bad actors trying to access their accounts. Our greatest challenge is protecting customer trust without unjustly harming good customers. To strike the right balance, we invest in mechanisms which allow us to accurately identify and mitigate risk, and to quickly correct and learn from our mistakes. This strategy includes continuously evolving enforcement policies, iterating our Machine Learning risk models, and exercising high‐judgement decision‐making where we cannot apply automation.
US, CA, San Francisco
LOCATION: San Francisco, CAMULTIPLE POSITIONS AVAILABLE1. Analyze real user data (search query logs) using SQL or equivalent data query language.2. Train machine learning / deep learning based models using ML platforms and libraries such as Tensorflow, Pytorch, Pyspark etc.3. Apply natural language processing techniques to improve ranking of search results and develop new ranking features and techniques building upon the latest results from the academic research community4. Boost search conversion by classifying user search queries and recommending relevant content5. Contribute to operational excellence in search team's scientific features, constructively identifying inefficient processes and proposing solutions6. Experiment with different models, analyze results using statistical methods and iterate on improving the results7. Propose and validate hypotheses to direct our business and product road map. Work with engineers to make low latency model predictions and scale the throughput of the system.8. Design, develop, and implement production level code that serves millions of search requests. Own the full development cycle: design, development, impact assessment, A/B testing (including interpretation of results) and production deployment.9. Telecommuting benefits available#0000