Rishi Rajasekaran
Applied Scientist at Amazon specializing in recommender systems
- Role
- Applied Scientist at Amazon
- Location
- Sunnyvale, CA, US
- LinkedIn followers
- 500 followers
About Rishi Rajasekaran
I am an Applied Scientist at Amazon in Alexa AI. My key interests and competencies…
Experience
Applied Scientist
Aug 2022 — Present · Sunnyvale, CA, US
Applied Scientist in the GENIE team under the Devices (Alexa) org. I developed LLM-powered content recommendation models for ambient display content on Echo Show devices for Alexa+ to enable personalization based on user interests. Currently working on systems for enabling new Alexa+ features for proactive experiences. Ranking- Worked on ranking using LLMs, engagement modeling, similar content deduplication and diversification and designing LLM-as-a-Judge based evals for validating the performance of personalization models for customer interests which was launched as part of the new Alexa+ experience- In 2023, I developed a provider level impression allocation model for Ambient Display content recommendations that contributed to a 5% lift in voice and touch engagement in NA as well as offline experimentation on a neural ranker for ranking providers based on engagement. Conversational Ambiguity- In 2022, I worked on fine-tuning pretrained encoder models and transformer models for detecting conversational ambiguity from voice utterances that enabled Alexa to perform conversational error recovery (primarily in ASR and NLU) through clarification question asking- Helped launch models in new marketplaces worldwide (North America, Europe, Asia) enabling conversational error recovery for Alexa customers across 5 languages- Instrumental in helping team attain 40% of 2022 KPI target of conversational defect detection in less than 4 months of joining through modeling work in detecting incomplete user utterances.
Education
University of California, Santa Cruz
Master of Science - MS, Computer Science
2019 — 2020
National Institute of Technology, Tiruchirappalli
Bachelor of Technology (B.Tech.), Computer Science
2013 — 2017
Skills
- Php
- Javascript
- C
- Programming
- Algorithms
- Web Development
- C++
- Html5
- Java
- Node.js
- Mysql
- Sql
- Linux
- Php Frameworks
- Angularjs
- Python
- Html
- Laravel
Find verified contacts for anyone on LinkedIn
Unifers gives sales teams verified emails and direct dials, enriched profiles, and outreach that lands in the inbox.
Free plan included · No credit card required
This profile is compiled from publicly available professional sources. Unifers is not affiliated with or endorsed by LinkedIn. Request removal of this profile.