$ whoami
> AI systems engineer. I build retrieval pipelines, autonomous agents,
> and the backend infrastructure that keeps them fast at production scale.
$ status --current
> Focused on RAG evaluation & agentic browser automation · CSE undergrad
2ND PLACE · IIT ROORKEE E-SUMMIT 2026 — 550+ teams
RAG + NLI engine that validates claims against 10,000+ document chunks and returns evidence-backed verdicts with a trust score, served through a real-time React dashboard.
91% NLI accuracy · +38% retrieval quality · <200ms latency · 10K+ chunks
Python · LangChain · SBERT · RoBERTa · React · PostgreSQL
SELECTED · T-HUB HYDERABAD INCUBATION — top 1% of 1,000+ teams
Agent that executes multi-step workflows across 100+ pages per session; concurrent Playwright pipelines cut execution time from 55 to 7 minutes with LLM + rule-based extraction.
87% faster execution · 50K+ records · 95% extraction accuracy
Python · Playwright · LangChain · AsyncIO
RAG EVALUATION FRAMEWORK
Coverage scoring and top-k analysis across 1,000+ queries — surfaces missing evidence (30–45% in audited pipelines) and cuts irrelevant retrieval noise by 28%, with explainable diagnostics.
1K+ queries audited · −28% retrieval noise · +60% debug speed
Python · FAISS · OpenAI · Streamlit
languages ── Python · TypeScript · C++ · JavaScript
ai / ml ── LangChain · SBERT · RoBERTa · FAISS · Pinecone · PyTorch
backend ── FastAPI · Flask · Node.js · PostgreSQL · Supabase
infra ── Docker · Azure · GitHub Actions · Linux
open source ── 15+ merged PRs in production codebases
algorithms ── 1,450+ problems · LeetCode / CodeChef
community ── Microsoft Learn Student Ambassador · 200+ students mentored
certified ── Microsoft AZ-900 / AI-900 / DP-900 · Oracle OCI · Google Cloud
open to backend / AI engineering internships — summer 2026