Indore, IN 22.72°N 75.86°E B.Tech CSE — Medicaps ’27 AI/ML Engineer Intern — GeeCom India · May 2026→present

Somya Sahu

AI/ML engineer and full-stack builder. I take systems the whole distance — agentic pipelines, real-time computer vision, CUDA kernels, pricing engines — and most of them were trained, profiled and benchmarked on one 6 GB RTX 3050.

Selected work 12somyasahu@gmail.com GitHub Résumé
Fig. 01 — Collatz total stopping time, consecutive n n = 0
0%
Form-classifier accuracy, GymGuard
0B/s
Integers verified on one RTX 3050
0
Pose samples recorded by hand
0+
Countries scored live in GEO·INTEL
0fps
Real-time inference, one webcam

Selected
systems

Five builds that carry numbers. Pick one — the sheet beside it holds the specs, the stack and the source.

Also shipped

Smaller tools, both live and open source.

Working
set

What I actually reach for. Highlighted entries are the ones I go deep on.

Head of Technology — Medicaps E-Sports Club

Tournaments
that scale

I run the technology behind large-scale sponsored tournaments: registration and seeding, bracket operations, live production, and the event-day infrastructure that has to stay up while a few hundred people are watching. Brands that backed those events:

Red BullAMDFormula 1JioGames
Open to AI/ML and systems work · internships · collaborations

Let’s build
something fast

12somyasahu@gmail.comEmail ↗ github.com/12somyasahuSource ↗ linkedin.com/in/12somyasahuLinkedIn ↗ RésuméOpen PDF ↗