There’s a specific feeling I’m chasing with this blog: the moment a confusing thing suddenly makes sense. Not the textbook version of the concept. The messy, actual process of being wrong, then less wrong, then occasionally right.
I’m Nishanth. I’ve spent a decade building ML systems in fintech, operations, consulting, marketing, engineering, and now insurance, and I’m still regularly confused by things I feel I should understand by now. This blog is where I work through that.
Short posts about things that confused me and how I eventually resolved them, or about things I’m intrigued by. Occasionally longer posts walking through experiments and implementations from notebooks.
ML, OMSCS coursework, hardware tinkering, and whatever else currently has my attention - watercolors, cycle mechanics, guitar. The domain doesn’t matter much. The learning posture does.
The approach is Feynman-ish: if I can’t explain it clearly, I probably don’t understand it yet.
I try to write at the level of someone who knows the domain but hasn’t encountered this specific nuance - usually me, six months ago.
Lead Data Scientist at OIP Insurtech, currently focused on an AI-driven insurance triage product - BoundAI.
Previously: ML frameworks in fintech at scale, NLP-driven contact centre analytics, market research, and more.
Education: M.S. in Computer Science at Georgia Tech (OMSCS) · B.E. in Electrical and Electronics Engineering.
Functional fitness, cycling, backpacking. Read a lot. Working through physical rehabilitation post traumatic injuries - swimming is part of the comeback.