Archive
Writing
Notes on agentic RL, video AI, and the path from research to production.
2025
Why Agentic RL Changes Everything
Feb 28
Most RL research treats the agent as a policy optimizing a fixed reward. Agentic RL is different — the agent reasons, plans, uses tools, retries.
The Hard Problems of RL for Video AI
Feb 20
Video generation models need evaluation beyond FID/FVD. Reward design, credit assignment, training instability, and the data flywheel cold start.
Building Reward Models That Actually See
Feb 10
The gap between “generates text well” and “evaluates video well” is enormous. From CLIP scores to learned reward models calibrated on expert data.
The AI Divide: How We Use ChatGPT vs. Claude
Sep 19
Comparative analysis of how users engage with ChatGPT and Claude, based on the latest AI research.
2022
Recent FAQs — DS & ML Job Seeking
Oct 17
How I chose DS/ML over SWE, interview prep, and reflections on working at Microsoft.