Ujwal Shyam Kantimohanthy
Internet engineer — networking, AI systems, space infrastructure · Currently in Rome, relocating to Paris (STATION F Landing Zone)
Most of what I work on comes down to the same question: what's actually holding this up, and can I trust it? I studied internet engineering because networks are one of the few systems you can watch fail in real time. I still think in that shape — trace the problem down to the layer underneath, then build there.
- Preparing to relocate to Paris for STATION F's Landing Zone (Aug 10 – Sep 10)
- A vertical slice through Computer Networking — one topic, DNS or HTTP, taken from source to real experiment to artifact
- 3D cell tracking for the Kaggle × ChanZuckerberg Biohub competition, due late September
Evidence over percentages
I don't put skill bars or proficiency scores on anything here. If I can't point to the artifact — code, a config, a diagram, a number someone else can check — I don't claim it.
Show what broke
Most write-ups skip the part where the plan didn't work. I try not to. The middle of a project is usually more useful than the pitch.
Build the layer underneath
Given a choice between the visible feature and the thing it depends on, I'll usually pick the dependency. Networks, ontologies, validators — unglamorous, but everything else stands on them.
One connected map
Math, electronics, networks, distributed systems, AI — I stopped treating these as separate subjects a while ago. They're one map at different zoom levels.