Hello! I’m Pararth Shah, a Senior Staff ML Engineer at Google in SF Bay Area.
I find joy in solving important, hard problems with talented, low-ego people, and in the process creating wealth, meaning, and fun for everyone involved. My guiding principle: per aspera ad astra (through hardships to the stars).
Professional Journey
My journey in the tech world began at the IIT Bombay, where I earned a Bachelor of Technology in Computer Science & Engineering. I first got introduced with Artificial Intelligence and Machine Learning during my time there, and conducted research with some amazing professors (Dr. Pushpak Bhattacharya, Dr. Sunita Sarawagi, Dr. S. Sudarshan, Dr. Ganesh Ramakrishnan, and others).
Then I got a Master of Science in Computer Science from Stanford University. The highlights of my time there include taking Dr. Andrej Karpathy’s and Dr. Fei-Fei Liu’s inaugral class on neural networks for computer vision, taking Dr. Andrew Ng’s famed Machine Learning class that felt like attending a concert, working in Dr. Jure Leskovec’s lab on graph databases, sitting in the front row through Sam Altman’s class on How To Start a Startup, and hustling through Steve Blank’s Lean Launchpad course to get to product market fit within one academic quarter.
Since then I’ve had the pleasure of working with some of the most innovative companies in tech, including:
Google: Led teams to develop the agentic framework for Gemini and rebuild the Google Assistant with Large Language Models (LLMs). I’ve managed teams to upgrade Google Assistant's NLP stack, shipping state-of-the-art NLP and Conversational AI techniques to enhance human-level interactions.
Facebook: Developing voice-enabled conversational assistants for the “metaverse” at Facebook Reality Labs, the voice assistant for Portal, Oculus and Rayban Stories, and Messenger Assistant.
Sourcegraph: Was among the first ten hires, contributed to building a revolutionary software development platform.
Publications and Contributions
I have contributed to the academic community through several publications (with 2500+ citations), including:
Opendialkg: Explainable conversational reasoning with attention-based walks over knowledge graphs
Bootstrapping a neural conversational agent with dialogue self-play, crowdsourcing and on-line reinforcement learning
Interactive Reinforcement Learning for Task-Oriented Dialogue Management
Ringo: Interactive Graph Analytics on Big-Memory Machines
Contact
Feel free to connect with me on LinkedIn.
Subscribe to get full access to the newsletter and website. Never miss an update.
