Summary
Sachit Menon is a PhD candidate and graduate student researcher at Columbia University with 9+ years of ML and computer vision research experience focused on multimodal foundation models, agentic reasoning, and inference-time methods. He has bridged academia and industry—contributing to Gemini 2.5 and publishing a video reasoning benchmark at ICCV 2025 during a research visit to Google DeepMind, and developing multimodal LLM+diffusion systems at Meta (CVPR 2024). His work spans training, post-training (SFT/RL), tool-using LLM agents that generate code for vision tools (ViperGPT), and designing evaluations and guardrails to improve reliability and safety. Sachit combines large-scale distributed experimentation with careful dataset curation and failure-mode analysis, having worked with annotation pipelines and metrics for video understanding. He’s on the job market for Research Scientist roles in vision-language models, multimodal evaluation, RL/post-training, and trustworthy AI. Notably, his projects emphasize practical agent workflows—programmatic tool orchestration and inference-time compute scaling—to make multimodal reasoning robust in real-world settings.
9 years of coding experience
5 years of employment as a software developer
High School Diploma, High School Diploma at Texas Academy of Mathematics and Science, University of North Texas
Bachelor of Science - BS, Mathematics and Computer Science, Bachelor of Science - BS, Mathematics and Computer Science at Duke University
Doctor of Philosophy - PhD, Computer Science, Doctor of Philosophy - PhD, Computer Science at Columbia University
French, Japanese