Blog Posts
Research notes, project write-ups, and shorter pieces from ongoing work.
Robot Learning: A Mathematical Guide to Training Policies (in progress work)
A guided look at imitation error, compounding, DAgger, stability, and the limits of continuous-action learning guarantees.Does Diffusion Actually Capture Multi-Modal Behavior?
Re-running the Much Ado About Noising myth-buster tests with diffusion policies on genuinely multi-modal robot data. Also available as a thread on X.My Punk Advice for Undergrads
A departure from the usual research notes — advice I felt compelled to consolidate from my undergrad years.Notes on Modafinil (Written on 100mg)
Ever since humans sat down and tried to work for long hours, we’ve struggled with it.How Fine-Tuning Rewires AI: A Mechanistic Interpretability Study
An early experiment studying training dynamics through a mechanistic-interpretability lens.We Tried to Make LLM-GP Evolution Better
Notes from an attempt to improve LLM-guided genetic programming.