Interspeech
This year's Interspeech will be held in Graz, Austria, whose famed clock tower was built in the mid-1500s
Photo courtesy of Getty Images

The 16 Alexa-Related Papers at This Year’s Interspeech

At next week’s Interspeech, the largest conference on the science and technology of spoken-language processing, Alexa researchers have 16 papers, which span the five core areas of Alexa functionality: device activation, or recognizing speech intended for Alexa and other audio events that require processing; automatic speech recognition (ASR), or converting the speech signal into text; natural-language understanding, or determining the meaning of customer utterances; dialogue management, or handling multiturn conversational exchanges; and text-to-speech, or generating natural-sounding synthetic speech to convey Alexa’s responses. Two of the papers are also more-general explorations of topics in machine learning.

Device Activation

Model Compression on Acoustic Event Detection with Quantized Distillation
Bowen Shi, Ming Sun, Chieh-Chi Kao, Viktor Rozgic, Spyros Matsoukas, Chao Wang

The researchers combine two techniques to shrink neural networks trained to detect sounds by 88%, with no loss in accuracy. One technique, distillation, involves using a large, powerful model to train a leaner, more-efficient one. The other technique, quantization, involves using a fixed number of values to approximate a larger range of values.

Sub-band Convolutional Neural Networks for Small-footprint Spoken Term Classification
Chieh-Chi Kao, Ming Sun, Yixin Gao, Shiv Vitaladevuni, Chao Wang

Convolutional neural nets (CNNs) were originally designed to look for the same patterns in every block of pixels in a digital image. But they can also be applied to acoustic signals, which can be represented as two-dimensional mappings of time against frequency-based “features”. By restricting an audio-processing CNN’s search only to the feature ranges where a particular pattern is likely to occur, the researchers make it much more computationally efficient. This could make audio processing more practical for power-constrained devices.

A Study for Improving Device-Directed Speech Detection toward Frictionless Human-Machine Interaction
Che-Wei Huang, Roland Maas, Sri Harish Mallidi, Björn Hoffmeister

This paper is an update of prior work on detecting device-directed speech, or identifying utterances intended for Alexa. The researchers find that labeling dialogue turns (distinguishing initial utterances from subsequent utterances) and using signal representations based on Fourier transforms rather than mel-frequencies improve accuracy. They also find that, among the features extracted from speech recognizers that the system considers, confusion networks, which represent word probabilities at successive sentence positions, have the most predictive power.

Automatic Speech Recognition (ASR)

Acoustic Model Bootstrapping Using Semi-Supervised Learning
Langzhou Chen, Volker Leutnant

The researchers propose a method for selecting machine-labeled utterances for semi-supervised training of an acoustic model, the component of an ASR system that takes an acoustic signal as input. First, for each training sample, the system uses the existing acoustic model to identify the two most probable word-level interpretations of the signal at each position in the sentence. Then it finds examples in the training data that either support or contradict those probability estimates, which it uses to adjust the uncertainty of the ASR output. Samples that yield significant reductions in uncertainty are preferentially selected for training.

Improving ASR Confidence Scores for Alexa Using Acoustic and Hypothesis Embeddings
Prakhar Swarup, Roland Maas, Sri Garimella, Sri Harish Mallidi, Björn Hoffmeister

Speech recognizers assign probabilities to different interpretations of acoustic signals, and these probabilities can serve as inputs to a machine learning model that assesses the recognizer’s confidence in its classifications. The resulting confidence scores can be useful to other applications, such as systems that select machine-labeled training data for semi-supervised learning. The researchers append embeddings — fixed-length vector representations — of both the raw acoustic input and the speech recognizer’s best estimate of the word sequence to the inputs to a confidence-scoring network. The result: a 6.5% reduction in equal-error rate (the error rate that results when the false-negative and false-positive rates are set as equal).

Multi-Dialect Acoustic Modeling Using Phone Mapping and Online I-Vectors
Harish Arsikere, Ashtosh Sapru, Sri Garimella

Multi-dialect acoustic models, which help convert multi-dialect speech signals to words, are typically neural networks trained on pooled multi-dialect data, with separate output layers for each dialect. The researchers show that mapping the phones — the smallest phonetic units of speech — of each dialect to those of the others offers comparable results with shorter training times and better parameter sharing. They also show that recognition accuracy can be improved by adapting multi-dialect acoustic models, on the fly, to a target speaker.

Neural Machine Translation for Multilingual Grapheme-to-Phoneme Conversion
Alex Sokolov, Tracy Rohlin, Ariya Rastrow

Grapheme-to-phoneme models, which translate written words into their phonetic equivalents (“echo” to “E k oU”), enable speech recognizers to handle words they haven’t seen before. The researchers train a single neural model to handle grapheme-to-phoneme conversion in 18 languages. The results are comparable to those of state-of-the-art single-language models for languages with abundant training data and better for languages with sparse data. Multilingual models are more flexible and easier to maintain in production environments.

Scalable Multi Corpora Neural Language Models for ASR
Anirudh Raju, Denis Filimonov, Gautam Tiwari, Guitang Lan, Ariya Rastrow

Language models, which compute the probability of a given sequence of words, help distinguish between different interpretations of speech signals. Neural language models promise greater accuracy than existing models, but they’re difficult to incorporate into real-time speech recognition systems. The researchers describe several techniques to make neural language models practical, from a technique for weighting training samples from out-of-domain data sets to noise contrastive estimation, which turns the calculation of massive probability distributions into simple binary decisions.

Natural-Language Understanding

Neural Named Entity Recognition from Subword Units
Abdalghani Abujabal, Judith Gaspers

Named-entity recognition is crucial to voice-controlled systems — as when you tell Alexa “Play ‘Spirit’ by Beyoncé”. A neural network that recognizes named entities typically has dedicated input channels for every word in its vocabulary. This has two drawbacks: (1) the network grows extremely large, which makes it slower and more memory intensive, and (2) it has trouble handling unfamiliar words. The researchers trained a named-entity recognizer that instead takes subword units — characters, phonemes, and bytes — as inputs. It offers comparable performance with a vocabulary of only 332 subwords, versus 74,000-odd words.

Dialogue Management

HyST: A Hybrid Approach for Flexible and Accurate Dialogue State Tracking
Rahul Goel, Shachi Paul, Dilek Hakkani-Tür

Dialogue-based computer systems need to track “slots” — types of entities mentioned in conversation, such as movie names — and their values — such as Avengers: Endgame. Training a machine learning system to decide whether to pull candidate slot values from prior conversation or compute a distribution over all possible slot values improves slot-tracking accuracy by 24% over the best-performing previous system.

Towards Universal Dialogue Act Tagging for Task-Oriented Dialogues
Shachi Paul, Rahul Goel, Dilek Hakkani-Tür

Dialogue-based computer systems typically classify utterances by “dialogue act” — such as requesting, informing, and denying — as a way of gauging progress toward a conversational goal. As a first step in developing a system that will automatically label dialogue acts in human-human conversations (to, in turn, train a dialogue-act classifier), the researchers create a “universal tagging scheme” for dialogue acts. They use this scheme to reconcile the disparate tags used in different data sets.

Topical-Chat: Towards Knowledge-Grounded Open-Domain Conversations
Karthik Gopalakrishnan, Behnam Hedayatnia, Qinlang Chen, Anna Gottardi, Sanjeev Kwatra, Anu Venkatesh, Raefer Gabriel, Dilek Hakkani-Tür

The researchers report a new data set, which grew out of the Alexa Prize competition and is intended to advance research on AI agents that engage in social conversations. Pairs of workers recruited through Mechanical Turk were given information on topics that arose frequently during Alexa Prize interactions and asked to converse about them, documenting the sources of their factual assertions. The researchers used the resulting data set to train a knowledge-grounded response generation network, and they report automated and human evaluation results as state-of-the-art baselines.

Text-to-Speech

Towards Achieving Robust Universal Neural Vocoding
Jaime Lorenzo Trueba, Thomas Drugman, Javier Latorre, Thomas Merritt, Bartosz Putrycz, Roberto Barra-Chicote, Alexis Moinet, Vatsal Aggarwal

A vocoder is the component of a speech synthesizer that takes the frequency-spectrum snapshots generated by other components and fills in the information necessary to convert them to audio. The researchers trained a neural-network-based vocoder using data from 74 speakers of both genders in 17 languages. The resulting “universal vocoder” outperformed speaker-specific vocoders, even on speakers and languages it had never encountered before and unusual tasks such as synthesized singing.

Fine-Grained Robust Prosody Transfer for Single-Speaker Neural Text-to-Speech
Viacheslav Klimkov, Srikanth Ronanki, Jonas Rohnke, Thomas Drugman

The researchers present a new technique for transferring prosody (intonation, stress, and rhythm) from a recording to a synthesized voice, enabling the user to choose whose voice will read recorded content, with inflections preserved. Where earlier prosody transfer systems used spectrograms — frequency spectrum snapshots — as inputs, the researchers’ system uses easily normalized prosodic features extracted from the raw audio.

Machine Learning

Two Tiered Distributed Training Algorithm for Acoustic Modeling
Pranav Ladkat, Oleg Rybakov, Radhika Arava, Sree Hari Krishnan Parthasarathi,I-Fan Chen, Nikko Strom

When neural networks are trained on large data sets, the training needs to be distributed, or broken up across multiple processors. A novel combination of two state-of-the-art distributed-learning algorithms — GTC and BMUF — achieves both higher accuracy and more-efficient training then either, when learning is distributed to 128 parallel processors.

BMUF-GTC.gif._CB436386414_.gif
The researchers' new method splits distributed processors into groups, and within each group, the processors use the highly accurate GTC method to synchronize their models. At regular intervals, designated representatives from all the groups use a different method — BMUF — to share their models and update them accordingly. Finally, each representative broadcasts its updated model to the rest of its group.
Animation by Nick Little

One-vs-All Models for Asynchronous Training: An Empirical Analysis
Rahul Gupta, Aman Alok, Shankar Ananthakrishnan

A neural network can be trained to perform multiple classifications at once: it might recognize multiple objects in an image, or assign multiple topic categories to a single news article. An alternative is to train a separate “one-versus-all” (OVA) classifier for each category, which classifies data as either in the category or out of it. The advantage of this approach is that each OVA classifier can be re-trained separately as new data becomes available. The researchers present a new metric that enables comparison of multiclass and OVA strategies, to help data scientists determine which is more useful for a given application.

About the Author
Larry Hardesty is a science writer at Amazon. Previously, he was managing editor of the Boston Book Review, a senior editor at MIT Technology Review, and the computer science writer at the MIT News Office.

Work with us

See More Jobs
US, CA, East Palo Alto
Have you ever wanted to work on state of the art computer vision, natural language processing and applied machine learning that will make a lasting impact on society?We are looking for brilliant Machine Learning Scientists who have the passion to tackle tough problems by bringing cutting edge machine learning technologies into the hands of our customers!As an Applied Scientist on the Amazon AI Team, you will research, experiment, iterate and publish on cutting-edge AI technologies. You'll also work closely with engineers and product managers to design and develop fast, efficient, and highly scalable deep learning algorithms that are applied to challenging every-day use case problems.
US, CA, Sunnyvale
Interested in Amazon Echo? We’re building the speech and language solutions behind Amazon Echo and other Amazon products and services. Come join us!As a Applied Scientist, you will be responsible for data-driven improvements to our spoken language understanding models. Your work will directly impact our customers in the form of products and services that make use of speech and language technology.You will:· Ensure data quality throughout all stages of acquisition and processing, including such areas as data sourcing/collection, ground truth generation, normalization, transformation, cross-lingual alignment/mapping, etc.· Clean, analyze and select data to gauge customer experience· Build and release models that elevate the customer experience and track impact over time· Collaborate with colleagues from multidisciplinary science, engineering and business backgrounds.· Present proposals and results in a clear manner backed by data and coupled with actionable conclusions to drive business decisions· Work with engineers to develop efficient data querying and modeling infrastructure for both offline and online use cases
US, WA, Seattle
Do you want to join Alexa AI -- the science team behind Amazon’s intelligence voice assistance system? Do you want to utilize cutting-edge deep-learning and machine learning algorithms to delight millions of Alexa users around the world?If your answers to these questions are “yes”, then come join us at the Alexa Artificial Intelligence team, which is in charge of improving Alexa user satisfaction through real-time metrics monitoring and continuous closed-loop learning. The team owns the modules that reduce user perceived defects and frictions through utterance reformulation, contextual and personalized hypothesis ranking.With the Alexa Artificial Intelligence team, you will be working alongside a team of experienced machine/deep learning scientists and engineers to create data driven machine learning models and solutions on tasks such as sequence-to-sequence query reformulation, graph feature embedding, personalized ranking, etc..You will be expected to:· Analyze, understand, and model user-behavior and the user-experience based on large scale data, to detect key factors causing satisfaction and dissatisfaction (SAT/DSAT).· Build and measure novel online & offline metrics for personal digital assistants and user scenarios, on diverse devices and endpoints· Create and innovate deep learning and/or machine learning based algorithms for utterance reformulation and contextual hypothesis ranking to reduce user dissatisfaction in various scenarios;· Perform model/data analysis and monitor user-experienced based metrics through online A/B testing;· Research and implement novel machine learning and deep learning algorithms and models.
US, CA, Palo Alto
Amazon is building a world class advertising business and defining and delivering a collection of self-service performance advertising products that drive discovery and sales of merchandise. Our products are strategically important to our Retail and Marketplace businesses driving long term growth. We deliver billions of ad impressions and millions of clicks daily and are breaking fresh ground to create world-class products. We are highly motivated, collaborative and fun-loving with an entrepreneurial spirit and bias for action. With a broad mandate to experiment and innovate, we are growing at an unprecedented rate with a seemingly endless range of new opportunities.We are looking for a Data Scientist for the Sponsored Products Search Supply team. Sponsored Products helps merchants, retail vendors, and brand owners succeed via native advertising that grows incremental sales of their products sold through Amazon. The Search Supply team works on complex engineering, optimization, econometric, and user-experience problems in order to deliver relevant product ads on Amazon search pages world-wide. Using a combination of big-data analytics, machine learning and experimentation, our scientists develop insights and optimizations that enable the monetization of Amazon online and mobile search properties while enhancing the experience of Amazon shoppers.The ideal candidate will have deep knowledge of data analysis/data science analysis ana and expertise with causal inferencing, A/B experiments, multi-variate testing , proven track-record to apply science knowledge to solve real-world problem from end-to-end , and ability to communicate data insights and scientific vision to a wide audience.As a Data Scientist, you will:· Perform hands-on analysis with very large data sets to develop insights that increase traffic monetization and merchandise sales without compromising shopper experience.· Work closely with data engineers to build data pipelines, maintain data lake to ensure the quality and timeliness of data management.· Design and run A/B experiments that affect hundreds of millions of customers, evaluate the impact of your optimizations and communicate your results to various business stakeholders.· Work with scientists and economists to model the interaction between organic sales and sponsored content and to further evolve Amazon's marketplace.· Establish scalable, efficient, automated processes for large-scale data exploratory analysis, machine-learning model creation and validation, and experiment insights generation.Amazon is an Equal Opportunity-Affirmative Action Employer – Minority / Female / Disability / Veteran / Gender Identity / Sexual Orientation.
DE, BE, Berlin
You are interested in Amazon Echo? Come work on it. We, Alexa's Applied Modeling & Data Science (AMDS) team in Berlin, are building the speech and language solutions for German and French behind Amazon Echo and other Amazon products and services.We are seeking an experienced, flexible and detail-oriented Senior Data Scientist that is passionate about working with language data. You will be responsible for data-driven improvements to our natural language understanding models, i.e., your work will have a direct impact on our customers.You will:· Ensure data quality throughout all stages of processing, including such areas as data sourcing/collection, ground truth generation, normalization, transformation, cross-lingual alignment/mapping, etc.· Continuously evaluate data analysis tools and processes and offer solutions to ensure they are efficient, high-quality and scalable· Lead and test tooling developments and pilot processes to support expansion to new data areas· Clean, analyze and select data to achieve goals· Present proposals and results in a clear manner backed by data and coupled with actionable conclusions· Closely collaborate with colleagues from science, engineering and business backgrounds
IN, KA, Bangalore
Do you want to join an innovative team of scientists who use machine learning and statistical techniques to create state-of-the-art solutions for providing better value to Amazon’s customers? Do you want to build and deploy advanced algorithmic systems that help optimize millions of transactions every day? Are you excited by the prospect of analyzing and modeling terabytes of data to solve real world problems? Do you like to own end-to-end business problems/metrics and directly impact the profitability of the company? Do you like to innovate and simplify? If yes, then you may be a great fit to join the Machine Learning and Data Sciences team for India Consumer Businesses.If you have an entrepreneurial spirit, know how to deliver, love to work with data, are deeply technical, highly innovative and long for the opportunity to build solutions to challenging problems that directly impact the company's bottom-line, we want to talk to you.Major responsibilities· Use machine learning and analytical techniques to create scalable solutions for business problems· Analyze and extract relevant information from large amounts of Amazon’s historical business data to help automate and optimize key processes· Design, development, evaluate and deploy innovative and highly scalable models for predictive learning· Research and implement novel machine learning and statistical approaches· Work closely with software engineering teams to drive real-time model implementations and new feature creations· Work closely with business owners and operations staff to optimize various business operations· Establish scalable, efficient, automated processes for large scale data analyses, model development, model validation and model implementation· Mentor other scientists and engineers in the use of ML techniques
IN, KA, Bangalore
Do you want to join an innovative team of scientists who use machine learning and statistical techniques to create state-of-the-art solutions for providing better value to Amazon’s customers? Do you want to build and deploy advanced algorithmic systems that help optimize millions of transactions every day? Are you excited by the prospect of analyzing and modeling terabytes of data to solve real world problems? Do you like to own end-to-end business problems/metrics and directly impact the profitability of the company? Do you like to innovate and simplify? If yes, then you may be a great fit to join the Machine Learning and Data Sciences team for India Consumer Businesses.If you have an entrepreneurial spirit, know how to deliver, love to work with data, are deeply technical, highly innovative and long for the opportunity to build solutions to challenging problems that directly impact the company's bottom-line, we want to talk to you.Major responsibilities· Use machine learning and analytical techniques to create scalable solutions for business problems· Analyze and extract relevant information from large amounts of Amazon’s historical business data to help automate and optimize key processes· Design, development, evaluate and deploy innovative and highly scalable models for predictive learning· Research and implement novel machine learning and statistical approaches· Work closely with software engineering teams to drive real-time model implementations and new feature creations· Work closely with business owners and operations staff to optimize various business operations· Establish scalable, efficient, automated processes for large scale data analyses, model development, model validation and model implementation· Mentor other scientists and engineers in the use of ML techniques
DE, BE, Berlin
Amazon AI is looking for world class scientists to join its AI Lab. This group is entrusted with developing core machine learning algorithms for AWS primarily around the time series domain. As a part of the AI Lab you will invent, implement, and deploy state of the art machine learning algorithms and systems. You will build prototypes and explore novel solutions to new problems at scale. You will interact closely with our customers and with the academic community.You will be at the heart of a growing and exciting focus area for AWS. We use both deep learning and probabilistic methods to solve exciting problems which require new methods more often than not.
US, WA, Seattle
Amazon.com strives to be Earth's most customer-centric company where people can find and discover anything they want to buy online. We hire the world's brightest minds, offering them a fast paced, technologically sophisticated and friendly work environment.This role will play an instrumental role in building, maintaining, and expanding core economic demand models for Amazon Devices. Using a conjoint-based economic demand application, you will simulate new products' impact on market segment share even before our customers ever see them on Amazon.com. You will experiment with constructing solutions to Device-specific business problems and help drive the scientific roadmap of our demand modeling software now used in other areas of Amazon. You will build econometric models, using our world class data systems, and apply economic theory to solve business problems in a fast moving environment.In addition, you will support up to SVP-level decision meetings around approving confidential funding requests (PRFAQs) for brand new devices and services, build decisions around how many hardware devices to build prior to receiving any customer signal, and pricing decisions around how to price and promote products and services. You will leverage Science and Tools produced by the Device Economics team such as conjoint demand models to produce these recommendations. As part of the stakeholder-facing arm of the team, you will own relationships with decision makers to help improve the customer experience by making the decisions that impact them more data and Science-driven. In parallel, you will work with Scientists, Economists, Product Managers, and Software Developers to provide meaningful feedback about stakeholder problems to inform business solutions and increase the velocity, quality, and scope behind our recommendations. In parallel we will invest in your development to pursue your career goals.
US, WA, Seattle
As the most customer-centric company on earth, Amazon is leading the way in Customer Service (CS) and recognized as #1 in the world. The Worldwide Defect Elimination (WWDE) organization in CS drives Amazon towards the creation of a defect-free customer experience by building technology and mechanisms to detect, escalate, and eliminate customer experience defects before they impact customers. To achieve this, we must understand what our customers are saying through different customer feedback sources including CS contacts, product reviews, return comments, and social media posts. Customer feedback is often in free form language and unstructured formats, requiring advanced science and engineering to derive insights.We are looking for talented Machine Learning (ML) scientists to build innovative and scalable models using state-of-the-art ML and Natural Language Processing (NLP) techniques to help Amazon understand Voice of the Customer and provide the perfect customer experience. Are you passionate about solving real world problems by analyzing and modeling millions of customer feedback data points every day and creating cutting-edge algorithms? Do you enjoy working on challenging problems, and implementing/deploying your ideas at scale? Are you capable of building end-to-end solutions that deliver significant benefits to the business in a fast pace environment? Will you bring innovation and simplicity to the team? If so, you are invited to join the WWDE Science Team.Major responsibilities:· Understand business challenges as well as identify gaps and new opportunities by analyzing and extracting relevant information from large amounts of both structured and unstructured data· Use statistics, NLP and deep learning techniques to create scalable, efficient, and automated solutions for business problems· Collaborate with tech and product teams on model building strategies and model experiments, implementation and continuous improvement· Improve the scalability, efficiency and automation of large-scale data analytics, model training, deployment and serving· Research and implement novel ML and statistical approachesAmazon is an Equal Opportunity-Affirmative Action Employer - Minority / Female / Disability / Veteran / Gender Identity / Sexual Orientation /Age
US, WA, Seattle
As the most customer-centric company on earth, Amazon is leading the way in Customer Service (CS) and recognized as #1 in the world. The Worldwide Defect Elimination (WWDE) organization in CS drives Amazon towards the creation of a defect-free customer experience by building technology and mechanisms to detect, escalate, and eliminate customer experience defects before they impact customers. To achieve this, we must understand what our customers are saying through different customer feedback sources including CS contacts, product reviews, return comments, and social media posts. Customer feedback is often in free form language and unstructured formats, requiring advanced science and engineering to derive insights.We are looking for talented Machine Learning (ML) scientists to build innovative and scalable models using state-of-the-art ML and Natural Language Processing (NLP) techniques to help Amazon understand Voice of the Customer and provide the perfect customer experience. Are you passionate about solving real world problems by analyzing and modeling millions of customer feedback data points every day and creating cutting-edge algorithms? Do you enjoy working on challenging problems, and implementing/deploying your ideas at scale? Are you capable of building end-to-end solutions that deliver significant benefits to the business in a fast pace environment? Will you bring innovation and simplicity to the team? If so, you are invited to join the WWDE Science Team.Major responsibilities:· Understand business challenges as well as identify gaps and new opportunities by analyzing and extracting relevant information from large amounts of both structured and unstructured data· Take the lead on medium-to-large business goals, and work on large-scale scientific projects, systems and processes· Use statistics, NLP and deep learning techniques to create scalable, efficient, and automated solutions for business problems· Collaborate with tech and product teams on model building strategies and model experiments, implementation and continuous improvement· Prototype and implement novel ML algorithms and statistical approaches· Coach and mentor junior scientistsAmazon is an Equal Opportunity-Affirmative Action Employer - Minority / Female / Disability / Veteran / Gender Identity / Sexual Orientation /Age
US, WA, Seattle
Are you looking to solve exciting Fraud challenges in the cloud? Do you want to work on a Cloud business that is growing multiple fold annually? If so, we want to talk to you! We are seeking a talented Data Scientist to help us evolve our processes and to improve customer experience across the globe.The mission of the Fraud Prevention Team is to keep the AWS platform a safe and trusted place for our customers and partners. We achieve this goal by identifying and preventing fraud and abuse for all AWS services, worldwide. Every day fraudsters attempt to steal our services. We use extensive data and machine learning models to build software solutions to detect and prevent such activities.As a Data Scientist, you will work directly with Research Scientists and Software Development Engineers to monitor the flavor/ trend of fraud worldwide and create appropriate actions to respond in a collaborative environment. There are no walls, and success is determined by your ability to dive deep, and understand the subtle demands new and complex services will place upon systems and teams.What would I do?· · Lead day-to-day review of emerging fraud threats and work with the broader team to drive responses to combat them· Manage your own process: identify and execute on high impact projects, triage external requests, and make sure you bring projects to conclusion in time for the results to be useful· Deep dive on the problems using SQL and scripting language like R, Python and internal dashboards and drive short term and long term solutions· Analyze data (past customer behavior, sales inputs, and other sources) to figure out trends, and output reports with clear recommendations.· Collaborate closely with the development team to recommend innovations based on data analysis· · Handle escalations from AWS Management for Fraud and related activities· · Be the primary point of contact during fraud outbreaks – including analyses to identify responses and applying responses.· Apply Statistical and Machine Learning methods to specific business problems and data· · Drive the design and development of tools to aid in operations and automation, such as creation of right dashboards to improve monitoring fraud· · Collaborate in a fast paced environment with multiple teams and customers in a dynamic entrepreneurial organization
US, CA, Sunnyvale
Amazon Lab126 is an inventive research and development company that designs and engineers high-profile consumer electronics. Lab126 began in 2004 as a subsidiary of Amazon.com, Inc., originally creating the best-selling Kindle family of products. Since then, we have produced groundbreaking devices like Fire tablets, Fire TV and Amazon Echo. What will you help us create?The Role:You will be working with a unique and gifted team developing exciting products for consumers. The team is a multidisciplinary group of engineers and scientists engaged in a fast paced mission to deliver new products. The team faces a challenging task of balancing cost, schedule, and performance requirements. You should be comfortable collaborating in a fast-paced and often uncertain environment, and contributing to innovative solutions, while demonstrating leadership, technical competence, and meticulousness.You will be responsible for validating the health of the thermal hardware and software architecture through development and ramp of a product. You are responsible for approving the thermal design is robust and meets the intended product requirements and expectations. A successful candidate will be passionately biased towards hands-on hardware validation, associated test methods, development of innovative/creative test validation workflows, automation of test workloads and device health analysis, along with a demonstrated working understanding of device thermal design. You will also demonstrate a strong desire to learn from new challenges, and demonstrate well developed problem solving and communication skills to work within a highly interactive, fast paced and diverse team. You will also be responsible for communicating device health status with internal customers and partners, coaching and mentoring sister organizations in best known validation practices. Candidates must demonstrate an ability to integrate feedback and agility to respond to changing product contexts/requirements.In this role, you will:· Validate thermal design architectures against their ability to meet product commits – be the single voice of thermal health of a device through your dedicated empirical test validation efforts· Contribute towards development of device requirements, validation metrics and associated software and hardware definitions· Be well versed in empirical thermal data collection methods including (but not limited to) device thermal instrumentation, AC and DC power supplies, and wide range of data acquisition systems (power, performance and thermals)· Be well versed in software tools like Android ADB or Python or equivalent to standardize, automate and enhance test throughput and quality· Understand software framework to debug device software, and ensure timely communication of thermal validation and device health· Own driving future improvements of the thermal validation process· Use of programming languages like Python and Matlab for analytical/statistical analyses and automation
US, MA, North Reading
The Research and Advanced Development team at Amazon Robotics is seeking passionate, hands-on scientists to work on cutting edge algorithms to power automation in Amazon’s order fulfillment and transportation network. You will have the chance to solve complex research problems and see your solutions come to life in Amazon’s warehouses!Our multi-disciplinary team includes scientists with backgrounds in AI planning and scheduling, machine learning, operations research, heuristic search, and robotic grasping. We work on problems such as:- Dynamic allocation and scheduling of tasks to thousands of robots- Learning how to recognize and grasp all the products that Amazon sells- Controlling the movement of many robots in tight spaces- Co-design of robotic logistics processes and the algorithms to optimize themThe ideal candidate for this position will be familiar with algorithms in one or more of these areas at both the theoretical and implementation levels.
US, WA, Bellevue
Amazon Fulfillment Services is looking for a motivated individual with strong analytical skills and practical experience to join our Modeling and Optimization team. We are hiring specialists at all levels into our scientific team with expertise in machine learning, network and combinatorial optimization, algorithm design, and/or control theory.Because we are driven by faster delivery to customers, a more efficient supply chain network, and lower cost of operations, our main focus is in the development of analytical strategic models and automation tools fed by our massive amounts of available data. You will be responsible for building these models/tools that improve the economics of the North American, European, and Japanese fulfillment networks as Amazon increases the speed and decreases the cost to deliver products to customers. You will identify and evaluate opportunities to reduce variable costs by improving the fulfillment center processes, transportation operations and scheduling, and the execution to operational plans. You will also improve the efficiency of capital investment by helping the fulfillment centers to improve storage utilization and the effective use of automation. Finally, you will help create the metrics to quantify improvements to the fulfillment costs (e.g., transportation and labor costs) resulting from the application of these optimization models and tools.The ideal candidate will have the following skills:- Ph.D. in Operations Research, Computer Science, Applied Mathematics or a related field with publications in refereed academic journals.- Good communication skills with both technical and business people. Ability to speak at a level appropriate for the audience.- Experience writing scripts (Perl, Python, Ruby, Groovy) to manipulate data and developing software in traditional programming languages (C++, Java).- Experience designing simulation and optimization models in for business decisions (e.g., staff/job scheduling, network routing, facility location) and/or feedback and/or predictive control.- Experience in designing/implementing online algorithms and approximation schemes for hard optimization problems (e.g. dynamic resource allocation problems for fulfillment center processes).- Experience designing/implementing machine learning algorithms tailored to particular business needs and tested on large datasets.- The ability to implement models and tools through the use of high-level modeling languages (e.g. AMPL, Mosel, R, Matlab, Julia) is a plus.
US, VA, Arlington
Amazon’s global Finance Operations team is looking for an experienced Senior Data Scientist to join our fast paced stimulating environment, to help invent the future of Accounts Receivable with technology, and to turn big data into actionable insights.The mission of Global Accounts Receivable is to foster positive interaction with our customers and optimise Free Cash Flow by improving the scale, speed, accuracy and productivity of the order-to-cash (O2C) cycle. We design O2C processes and systems so as to simplify business interaction between our customers and Amazon in all channels we support.The charter of the nascent Data Science area is to maximize Amazon’s return on our receivables investment in terms of Free Cash Flow, and customer satisfaction. We accomplish this by applying advanced statistical methods and empirical analysis to predict and evaluate customer behaviour. We provide data-driven recommendations to senior business leaders to optimize the O2C cycle in terms of policies, process and systems.We are seeking to hire a Senior Data Scientist with strong leadership and communication skills to join our team.Once in the role, you will help us build the future roadmap of Data Science at Global Accounts Receivable. You will discover and define problems; your quantitative solutions will impact the core business of Amazon. You will analyze large amounts of business data and develop metrics, insights and predictions for decision making at the leadership and the daily process level.Whether predictive customer behaviour analysis, forecasting of payment risk and mitigation across our sales channels and geographies, the design of learning processes, Global Accounts Receivable offers a plethora of quantitative areas with cash flow generation opportunities at a global scale.You will act as a thought leader of a team of software engineers, business intelligence engineers and business teams, to build accurate predictive models and algorithms, and deploy automated software solutions to make data more actionable to manage our business at scale.You will play an active role in translating business and functional requirements into concrete deliverables and working closely with software development teams to put solutions into Production.We are building a new team and this is an opportunity for you to define the scientific vision for this space.This role is based in Amazon's HQ2 in Arlington, VA.Selected Responsibilities· Apply business judgement to identify opportunities and develop science strategies· Design and develop predictive systems pertaining to customers’ payment behaviour· Run observational studies addressing dunning strategies across our global geographies and business channels· Apply advanced statistical analysis in a multi-variate, sequential process environment in order to identify cash flow drivers· Apply statistical and/or machine learning know-how e.g. to optimise the collections activities of our workforce· Develop new data sources to enable statistical modelling and learning; continuously fine-tune data models· Design and utilise code (Python, R, Scala, etc.) as required· Formulate experiments to assess AR process strategies· Collaborate with engineering to build data, algorithms and models· Communicate scientific solutions and insights effectively to a senior leadership and non-scientific audience
US, MA, Cambridge
Amazon is looking for a passionate, talented, and inventive Scientist with a strong machine learning background to help build industry-leading Speech and Language technology. Our mission is to push the envelope in Automatic Speech Recognition (ASR), Natural Language Understanding (NLU), and Audio Signal Processing, in order to provide the best-possible experience for our customers.As a Scientist, you will work with talented peers to develop novel algorithms and modeling techniques to advance the state of the art in spoken language understanding. Your work will directly impact our customers in the form of products and services that make use of speech and language technology. You will leverage Amazon’s heterogeneous data sources and large-scale computing resources to accelerate advances in spoken language understanding.We are hiring for Fairness and privacy initiatives in Alexa NLU.
US, WA, Seattle
We are a passionate team working to build a best-in-class healthcare product designed to make high-quality healthcare easy to access for our employee pilot program.As a Principal Research Scientist in Amazon Care, you will be responsible for setting our overall research direction, coordinating our research group, and working closely with stakeholders to understand their pain points.In this position, we are looking for a candidate whose experience blends hands-on clinical experience with a data science and machine learning background.
US, WA, Seattle
Prime Video is a global, high-growth business and a critical driver of Amazon Prime subscriptions, which contributes to customer loyalty and lifetime value. It is a digital video streaming and download service that offers Amazon customers the ability to rent, purchase or subscribe to a huge catalog of videos. The Prime Video and Music Economist team specializes in high-impact strategic analysis and modeling to improve the Prime Video flywheel. Now, the team is looking for a Senior Applied Scientist to develop and own new models individually and in partnership with other scientists on the team.As an Applied Scientist focusing on Amazon Video, you will be responsible for understanding the value that the business creates for our customers and to develop new, disruptive innovations to grow global Prime Video usage and customer value. This role requires an individual with strong quantitative modeling skills and the ability to apply statistical/machine learning, econometric, and experimental design methods to large amount of individual level data. The candidate should have strong communication skills, be able to work closely with stakeholders and translate data-driven findings into actionable insights. The successful candidate will be a self-starter comfortable with ambiguity, with strong attention to detail and ability to work in a fast-paced and ever-changing environment.The candidate's responsibilities will include:· Build scalable analytic solutions using state of the art tools based on large datasets· Build econometric models, conduct statistical/machine learning analyses, or design experiments to measure the value of the business and its many features· Partner closely with Business, Finance, Science, and Tech partners to build prototypes and implement production solutions· Develop and execute product work plans from concept to scalable production models· Write technical white papers and business-facing documents to clearly explain complex technical concepts to audiences with diverse business/scientific backgrounds
US, MA, Cambridge
We’re looking for a passionate, talented, and inventive scientist to help build industry-leading technologies in speech translation. Our team's mission is to enable Alexa to break down language barriers for our customers .As a Senior Applied Scientist on the Alexa Translations team, you will be responsible for developing novel algorithms that advance the state-of-the-art in speech translation, driving model and algorithmic improvements, formulating evaluation methodologies and for influencing design and architecture choices. You will leverage Amazon’s heterogeneous data sources and large-scale computing resources to build novel products and services that make use of speech and language technology. You will work in a hybrid, fast-paced organization where scientists and engineers work together and drive improvements to production. You will collaborate with and mentor other scientists to raise the bar of scientific research in Amazon.