Updated
I was born in India and now live in London.
I currently work at Springer Nature as Product Manager for SNAPP, launching initiatives that help researchers move from manuscript submission to peer-review, open-access funding, and payments.
Previously, I was a product lead at Noora Health, where I built digital healthcare products for patients, caregivers, and nurses across Bangladesh and Indonesia. Before that, I co-founded a health-tech hardware startup building ML-based hospital safety devices.
Alongside my professional work, I build agent-first products with LLMs and write about what I learn. Writing is how I think.
I remember being a big fan of Venn diagrams while studying set theory from mathematics, back in high school. I never touched any of this set theory, and all that jazz after that brief encounter. It was only recently while trying to navigate some complex decision making, that I realised how useful set theory is, as a mental model for visual reasoning: All it takes is a bunch of circles, that’s all...
When we see two events, A and B, our natural tendency is to place a succinct left-to-right arrow —>, in between two events A and B. It hardly crossed my mind sometimes that A and B could be feedback loops. A could effect B, and B could effect A, and the effect of B effecting A, could effect the effect of A effecting B…. . Why complicate things isn’t it? We will just leave it as it is, after all...
My agentic engineering workflow has changed in the recent past. The models are better, and there is much more freedom in choosing the harness, abilities, and actions you provide. - Pre-planning, pre-idea, pre-everything - The first chat with the agents - The first chat on the terminal - Setting up the Opencode environment - Doing light research, deep research - Model choices while working with...