Ankit Ranjan Das

Full Stack & AI Engineer

Featured Projects

OmniSearch

Architected a stable, end-to-end autonomous research platform utilizing Next.js and LangGraph, orchestrating a sequential 5-phase state machine (Planning, Crawling, Indexing, Retrieval, Publishing).

Next.js 14LangGraphFastAPIPostgreSQLRedisCeleryDocker Compose

View Project

DocuMind

Orchestrated a custom retrieval system combining Vector Search and BM25 Trigram algorithms using Reciprocal Rank Fusion (RRF); improving context retrieval accuracy by 40%.

Next.js 16RAGStripeInngest QueuesBM25

View Project

GitMetrix

Architected a hybrid RAG-based code intelligence engine analyzing GitHub repositories utilizing a Mixture-of-Experts LLM router (Cerebras, Groq, DeepSeek).

Next.js 16pgvectorSupabaseMoE RouterReact Flow

Technical Arsenal

Languages

C++PythonTypeScriptJavaScriptSQLHTML5 & CSS3

AI Engineering

LangGraphAgentic WorkflowsRAG PipelinesMoE RouterBAAI Cross-EncoderOllama

Frontend

Next.jsReact 19Tailwind CSS v4TanStack QueryFramer MotionReact Flow

Backend

FastAPINode.jsFlaskPostgreSQL (pgvector)SupabaseRedis

DevOps & Tools

Docker ComposeCeleryInngestGitWebSocketsVercel

Professional Journey

Tata Steel

Web Developer Intern

Internship

Engineered a real-time analysis backend using Python (Flask) & Scikit-Learn. Deployed Gradient Boosting Classifier (<100ms processing). Accelerated PostgreSQL query performance by 40% via compound indexing.