tldr: wannabe machine learning research and systems engineer who likes robotics applications
I am a software engineer at Google improving Cloud SDK build and release infrastructure. Previously, I graduated with a Computer Science degree from the University of Iowa. As a student, I interned at Google and improved Gemini Code Assist's Retrieval-Augmented Generation algorithms. I also worked on LLM-based computer-use agents and reinforcement learning.
My primary interest is in the nature of reality, mind, and Artificial Intelligence in particular. By AI I roughly mean computational systems that can efficiently learn a wide variety of useful tasks, drawing from Chollet's "On The Measure of Intelligence" and Hutter's AIXI. I prefer my work to make contact with reality and be guided by the creation of real value.
These interests and background converge to my goal: help make consumer technology based on these ideas (e.g. household robots) ubiquitous and "free" (as in freedom). This is a long-term goal and I don't know what path it will take, so I am always open to chat about any opportunities that may be tangentially related. Feel free to contact me@gautamsharda.com!
Relevant research / technical areas: robot learning, vision-language-action models, video-action models, world models, GPU programming, ML compilers, & hardware/software co-design.
Additional areas: general deep & reinforcement learning, robot design and manufacturing, neural network interpretability, PSI, non-gradient based methods, evolutionary algorithms, top-down program synthesis, automatic theorem proving.
"This is a long term, this is an infinite game. There is no 2 min, no end of quarter, there is no such thing." Forthrightness, substance, & depth.
Check out my projects.