Meet the Amazon science intern poster session winners
The company's first fully virtual class of interns got to showcase a digital copy of their poster, deliver a recorded presentation, and field questions about their projects.
This summer, more than 180 applied, research and data science graduate research interns had the opportunity to present a poster in Amazon’s first-ever virtual science intern poster session series. Amazon science interns typically present posters at the annual Graduate Research Colloquium event held in Seattle. However, due to the impact of COVID-19, this year’s event was virtual and took place over a six-week period, which allowed for inclusion of interns from across the globe.
Each week a group of interns had the opportunity to showcase a digital copy of their poster, deliver a recorded presentation, and field questions about their projects. More than 1,000 members of the Amazon science community were invited to rate, review and provide feedback on intern’s posters.
Below are the interns who had highest-rated poster presentations:
- Bhuvan Agrawal, Carnegie Mellon University: Learning cross-modal embeddings for end-to-end spoken language understanding
- Chuchu Cheng, Boston College: Aggregated time per package estimation
- Yaarit Even, Columbia University: Growing advertiser population with promotions: A causal inference approach
- Sanket Gaurav, University of Illinois at Chicago: Multi-task architecture and learning for controllers
- Ali Heydari, University of California, Merced: Deep-learned contextual representations of Amazon products for detail page ad selection and ranking
- Wenxiang Hu, University of Rochester: Improving performance of job recommender system based on classification and ranking model and click data
- Amit Kachroo, Oklahoma State University: Localization of objects (IoT)
- Ziren Lin, Columbia University: Coincidence estimation using machine learning
- Purvanshi Mehta, University of Rochester: Pre-training graph neural networks for NLP tasks
- Tejas Mehta, Carnegie Mellon University: Lumos sense interject - deep learning based anomaly detector for console sign-in detective control
- Harsh Parikh, Duke University: CREDENCE: credence for causal estimation
- Deepanshu Pariyani, University of Massachusetts Amherst: High dimensional anomaly detection for defect elimination
- Jingyi Xiao, University of California, Santa Barbara: Quantifying the Accessibility of Amazon Sites
- Xing Zhao, Texas A&M University: A biases analysis engine of SOTA recommender systems
- Kai Zhen, Indiana University Bloomington: Sparsification for on-device automatic speech recognition
- Liyuan Zheng, University of Washington: Product representation learning for substitute retrieval via heterogeneous graph neural networks
- Ivy Zhu, Ohio State University: IPC lab early experiment stopping