class Vetriselvan:
role = ["AI Researcher", "Systems Engineer"]
location = "Delhi NCR, India"
focus = ["Deep Learning", "Agentic AI", "LLM Systems", "Reinforcement Learning"]
building = [
"ARM-Gym - GRPO RL compiler optimizer for ARM AArch64",
"SRE-OpenEnv - agent training env on real Redis + PostgreSQL",
]
wins = [
"Top 4 · IIT Delhi ANRF AI-SE Hack 2026 - ResGRU-UNet PM2.5 forecasting",
"Finalist · Meta PyTorch OpenEnv Hack 2026 - SRE OpenEnv infrastructure agent",
"Top 5 · ANRF Phase 1 Online - #5 / 77 teams",
]
open_to = "Internships and new-grad roles · Aug 2026"
belief = "research the gap, build the fix, measure the delta"Deep Learning & ML
LLM / Agents
Languages & Web
| Project | What | Key result |
|---|---|---|
| PM2.5 Forecasting | ResGRU-UNet · gridded pollution 6h ahead | Top 4, score 0.8834 |
| ARM-Gym | GRPO RL compiler optimizer for ARM AArch64 | +14.5% over clang -O3 |
| SRE Incident Responder | OpenEnv · real Redis + PostgreSQL · 5 tasks | Finalist Meta PyTorch OpenEnv Hack |
| Credit Risk Scoring | LangGraph + RAG over RBI / SEBI guidelines | recall 0.41 → 0.65 |
| FinNuvora | AI finance · Plaid sync · Gemini doc intel | Full-stack · deployed |
