Bio

My name is Sanaz Bahargam. I am currently a tech lead working on Gemini’s training for new capabilities at DeepMind. Before Google DeepMind, I was an applied scientist and tech lead at Amazon Lab 126 (Alexa AI team) working on NLP, building large language models and LLM evalution. Before that I was at Twitter, working on Deep Learning, NLP and Learning to Rank problems with a focus on Search, Explore/Discovery, and Trends/Events products. During the last few years, I’ve been extensively working on Large Language Models, Transfer Learning, Few-Shot Learning, Multi-Task Learning, Representation Learning, Learning to Rank, Personalization/Recommendation and Generative AI for NLP.

I completed a Ph.D. degree at the Computer Science Department of Boston University in Machine Learning in 2017. During my Ph.D., I worked with Professor Evimaria Terzi, Professor Theodoros Lappas, and Professor  Vagelis Papalexakis. Before joining Data Mining and Machine Learning Group in 2015, I was a member of Networks Research Group and iBench Initiative working with Professor Azer Bestavros.

Besides work, I often play volley with friends, I enjoy hiking and amateur photography with my phone.

My Erdős number is 1. I have a publication with Erdős (not Paul, but with Dóra who happens to be a mathematician before entering the computer science field) :) For real my [Paul] Erdős number is 4.

Recent News

• Jan 27, 2022. I will be speaking about Deep Learning in Industry at Department of Computer Science - University of Colorado Boulder.
• Dec 04, 2021. I will be speaking about my work on using machine learning for Search and Recommendation at Debug Summit 2021.
• October 13, 2021. I will be speaking about my work on recommendations and ranking at Department of Computer Science - Saint Louis University.
• September 27, 2021. I will be speaking about my work on Alexa Conversations at NLP Summit 2021 this October.
• September 1, 2021. I will be speaking about my team’s work on task-based dialog system at Global #ArtificialIntelligence Conference this October. YouTube
• August 4, 2021. I’m happy to be a mentor for  #IGLUcontest @NeurIPSConf, Looking forward to helping the participants.
• February 2, 2020. I will be speaking at Lyft meetup about Trend/Event Detection and Recommendation @Twitter
• April 13, 2019. Our paper  The Guided TeamPartitioning Problem: Definition, Complexity, and Algorithm was accepted at the 12 International Conference on Educational Data Mining (EDM).
• November 27, 2018. My notes and summary of talks on the 2018 Conference on Digital Experimentation (CODE).
• October 25, 2018. Our paper  Team Formation Algorithm for Faultline Minimization was accepted at Expert Systems with Applications (ESWA).
• July 11, 2018. My notes on NAACL 2018.
• June 18, 2018. Our paper  Constrained Coupled Matrix-Tensor Factorization and its Application in Pattern and Topic Detection was accepted at ACM International Conference on Advances in Social Networks Analysis and Mining, 2018 (ASONAM).
• May 22, 2018. Our paper  Discovering Time-Evolving Topics of Varying Levels of Difficulty via Constrained Coupled Matrix-Tensor Factorization was accepted for poster presentation at the 4th International Conference on Computational Social Science (IC2S2).
• February 1, 2018. My notes on query understanding.
• July 24, 2017. I’m delighted to join Twitter as an ML engineer to work on NLP and ranking problems.
• June 15, 2017. I successfully defended my Ph.D. dissertation.