Rituraj

I work on test time interpretability, evaluation, and alignment for large language models. Down to play chess and talk about finance anytime.

Currently: building PRISM + evaluation systems.

Read PRISM

X logo GitHub logo YouTube logo LinkedIn logo

/recent

Mar 2026PRISM in 3 Parts: Scoring, Resampling, and Accept/Reject Filteringblog/Most parallel test-time scaling pipelines follow the same pattern: generate a pool of candidate solutions refine them over severa…[reasoning, inference, prm, search, systems]Mar 2026PRISM: Pushing the Frontier of Deep Think via Process Reward Model-Guided Inferenceresearch/TL;DR PRISM uses step-level verification (a PRM signal) to guide population refinement and solution aggregation in Deep Think sys…[research, reasoning, inference, PRM, deep-think]Mar 2026Diffusion Thinking, Autoregressive Speakingblog/I keep running into the same limitation with autoregressive language models. They commit too early. They can revise only by addin…[reasoning, inference, diffusion, autoregressive, systems]

/index

~/research2 itemsPapers and preprints~/projects1 itemPersonal and open source projects~/blog2 itemsBlog posts and articles

Press cmd+k to jump anywhere.

Last updated Mar 13, 2026