• 🤖 Building production Agentic AI systems — LangGraph agents, RAG pipelines, pgvector, multi-agent orchestration
• 💻 3.5+ years shipping full-stack web applications end-to-end — architecture → development → AWS deployment
• 🏆 Meta Hacker Cup Round 1 Qualifier · 2000+ competitive programming problems solved
• 🌍 Based in Ambikapur, Chhattisgarh, India

Objective-driven research agent: a LangGraph pipeline (planner → researcher → analyst → quality-check loop → report) where the user's goal steers every node prompt — so an acquisition brief and a partnership brief come out materially different. Streams node-by-node via SSE, Postgres-checkpointed state, grounded follow-up chat.

Tech: LangGraph · FastAPI · React · PostgreSQL · Tavily · Docker

Semantic anime recommender: offline FAISS index → embed query at request time → MongoDB metadata filters → LLM re-rank to ~12 picks, each with a one-line "why". AniList OAuth personalisation from watch history, LangGraph chat agent, weekly precision@k eval in CI — all on free-tier infrastructure.

Tech: Next.js · FastAPI · FAISS · MongoDB · LangGraph · LangSmith

AI content-creation platform for nonprofits. Production LangGraph agent (FastAPI) for multi-org social content generation, pgvector RAG knowledge base with HNSW indexing and hybrid keyword + vector search, Meta Graph API integration, multi-tenant architecture. Sole developer for 2.5+ years.

Tech: Next.js · Node.js · Python · LangGraph · PostgreSQL · pgvector · AWS

Multi-org social content generation agent powering Neki. A stateless LangGraph (FastAPI) pipeline that assembles bounded, org-scoped context from a pgvector knowledge base and drafts on-brand posts, with a resilient image-generation fallback ladder (gpt-image-1 → mini → dall-e-3 → dall-e-2 → PIL) and one-click publishing to Facebook, Instagram & LinkedIn via the Meta Graph API. Migrated from a Streamlit prototype to a decoupled Next.js + TanStack Query + FastAPI stack, with Chainlit as an optional debug harness.

Tech: LangGraph · FastAPI · Next.js · TanStack Query · pgvector · Meta Graph API

• ✅ Solved 2000+ coding problems across LeetCode, Codeforces & CodeChef
• 🏆 Meta Hacker Cup Round 1 Qualifier
• ⚡ LeetCode — codeantik_099

const ankit = {
  name: "Ankit Singh",
  roles: ["AI Engineer", "Full Stack Developer"],
  stack: {
    ai:        ["LangGraph", "LangChain", "RAG", "FAISS", "pgvector", "MCP"],
    backend:   ["Node.js", "FastAPI", "Express", "PostgreSQL", "MongoDB"],
    frontend:  ["Next.js", "React", "TypeScript", "Tailwind CSS"],
    infra:     ["AWS", "Docker", "Kubernetes", "CI/CD"],
  },
  currentFocus: [
    "Production agentic AI systems",
    "Multi-agent orchestration with LangGraph",
    "RAG & hybrid vector search pipelines",
    "Scalable full-stack SaaS architectures",
  ],
  funFact: "Solved 2000+ CP problems and still thinks in graphs 🧠",
};

⭐ From codeantik