___ ___ ___ ___ _ _ _____ _____ _ _ _____
/ _ \ | \/ | / _ \ | \ | | / ___|| _ | \ | |_ _|
/ /_\ \| . . |/ /_\ \| \| | \ `--. | | | | \| | | |
| _ || |\/| || _ || . ` | `--. \| | | | . ` | | |
| | | || | | || | | || |\ | /\__/ /\ \_/ / |\ |_| |_
\_| |_/\_| |_/\_| |_/\_| \_/ \____/ \___/\_| \_/\___/
Full-stack engineer moving deliberately into AI engineering β not "adding AI to CRUD apps," but building the systems underneath it: tool-calling agents, retrieval pipelines, and the backend infrastructure that makes them reliable in production.
I design the architecture and decisions; I use LLMs to help implement them β same discipline I bring to any other engineering tool.
π repo-rag β Tool-calling RAG agent over my own GitHub profile
A chat agent that answers questions like "has Aman used Kafka?" or "is CareerSpire finished?" in synthesized prose with citations β grounded in real repo content, not stale README claims.
The interesting decision: most personal-profile RAG demos batch-embed every repo upfront and do plain vector search. At the scale of one GitHub account, that's wasted work β most repos never get asked about. Instead the LLM decides what to explore at query time via tool calls, caches what it reads, and falls back across model tiers on failure.
Python FastAPI Tool-calling agents Model routing MCP
Ranked feed built on embeddings over event descriptions and user interaction signals β not a filtered list. The differentiator is the ranking model, not the CRUD around it.
Embeddings Ranking Node.js
Production-oriented interview practice platform: AI-generated questions and feedback, in-browser code execution, JWT auth with refresh tokens, rate limiting, 1000+ curated question bank.
Next.js Prisma Supabase LLM feedback loops
Generates architecture diagrams from natural-language prompts or a GitHub repo β landing page and design system live, dashboard/generation pipeline in progress.
Next.js Tailwind LLM-driven generation
Course marketplace with instructor/student dashboards, Stripe payments, and Supabase-backed auth β the backend-engineering fundamentals underneath the AI-first projects above.
Next.js Express Prisma Stripe
AI Engineering LLM tool-calling Β· RAG pipelines Β· Agent design Β· Prompt engineering Β· Model routing/fallback Β· MCP
Databases
Frontend
Cloud & Infra
Multi-agent systems Β· LangGraph Β· Model Context Protocol Β· Kubernetes Β· AI observability & evals