Abstract white particle field on black

Blog

Research, systems, and things learned while building.

I write about model research, post-training, inference, retrieval, and the engineering details that decide whether an AI system works outside a notebook.

Writing

4 posts
  1. Credit Assignment Is the Whole Game: A Modern Map of Reinforcement Learning from Pong to Post-TrainingA researcher-engineer guide to policy gradients, PPO, GRPO, RLOO, process rewards, and tool-use RL through the lens of credit assignment.
  2. When Your Multimodal RAG Team Punishes Good Retrieval: GraMRAG and Topology-Aware Policy OptimizationHow graph memory and topology-aware policy optimization assign credit across long-horizon multimodal retrieval trajectories.
  3. Multilingual tokenization without shortcutsDesign notes from building indicTok across 22 Indian languages and 12 scripts.
  4. Evidence-first RAG for high-stakes workflowsHow retrieval, citations, and human review fit together when plausible text is not enough.