Jharokha: Building a Photo Library That Doesn’t Own Your Photos
My rationale for building myself a photo library and some learnings from the process.
read more →writing & projects
Blog
Notes on machine learning, data science, research, and the occasional tangent.
Market Data Refinement: A Backtesting Stack to Simulate Strategies for Indian Equities
Details and learnings from my attempt at building a backtesting stack to simulate different strategies for picking Indian stocks.
read more →When Being Wrong Isn't Always the Same
On teaching decision trees to care about how they're wrong, not just whether they're wrong — and why that distinction quietly matters in reinforcement learning.
read more →Not Just Who to Call — Who to Call First
Predicting dropout is the easy part. The harder problem is deciding, with limited time and people, which of the hundreds of at-risk women will actually benefit from a health worker picking up the phone.
read more →A Woman Dies in Childbirth Every Fifteen Minutes
On working with ARMMAN and Google AI to predict which women were about to stop picking up the phone — and why that mattered more than it sounds.
read more →