Building production-grade AI systems: retrieval, agents, evals — shipped with clean APIs and real metrics.

---

⚡ TL;DR

• 🎓 B.Tech CSE @ VIT Bhopal (2023–2027) • CGPA 8.45 • internship-ready
• 🤖 AI focus: multi-agent pipelines, RAG with hybrid retrieval + evals, PEFT fine-tuning
• 🧩 Full-stack: Next.js UIs + FastAPI backends + Docker-deployed architecture
• 📊 I report real numbers — ROUGE, BERTScore, accuracy, latency — not just "built a model"

---

🧭 Current Stack

---

🧪 What I'm building / have shipped

• Multi-agent systems — LangGraph supervisor pattern, async job queues, LLM-driven routing, checkpointing
• RAG pipelines — hybrid search (BM25 + dense), RRF fusion, cross-encoder reranking, LLM-as-Judge evals
• Fine-tuning — PEFT/LoRA on FLAN-T5 and RoBERTa with reproducible metrics (ROUGE, BERTScore, F1)
• Production habits — Docker containers, FastAPI services, LangSmith tracing, MLflow/W&B experiment tracking

---

🧰 Tech Stack

Languages

ML / DL

LLM / Agent / Retrieval

MLOps / Observability

Backend / Full-Stack / DevTools

---

🎯 Fun Zone

---

📊 GitHub Stats

---

🐍 Contribution Snake

---

🎮 Off-screen side quests

• Gaming (PC / mobile)
• Music (EDM, lo‑fi, metal)
• Manhwa / webtoons
• Sci‑fi & thrillers

---

_{Built with curiosity, caffeine, and commits.}

---