Sep 2016 – May 2024 · 7 yrs 8 mos · Mountain View, California

• My work at Google was on applications of machine learning to natural language (using transformers). These are some of the hardest and most interesting problems in Computer Science and AI, and something I am very proud to have worked on.
• During my final year, I was leading the LLM modeling for Google's Generative Search (Search Generative Experience/AI Overviews), working across Search, Deepmind and Gemini teams.
• I also led multiple language generation efforts across different search products - question answering (featured snippets), related questions (people also ask), search web ranking, google autocomplete. I was awarded the Google Tech Impact Award for my work on natural language generation.
• Earlier, I led different projects in open domain question answering - misinformation detection (NLP techniques to detect unfounded, misleading or false content in Google Search), Factoid Question Answering (e.g., answering user queries like [what is the length of the statue of liberty's nose], [what tissue connects muscles to bones]), List Question Answering (e.g., [symptoms of kidney problems]), and Query Semantic Parsing (representing a natural language query in a machine understandable logical form).