Looking for a job

Janesh User

@janeshuser

Message

AI-focused Python developer building RAG applications, including a contract-to-billing reconciliation engine.

India

Message

I built a contract-to-billing reconciliation engine that uses RAG to extract contract terms and compare them with billing data. Its workflow combines Qdrant and Supabase pgvector retrieval with n8n automation and LLM fallback routing.

I also built a college FAQ assistant and a multimodal folder-search application. These projects use LangChain, embeddings, and vector retrieval to return context-grounded answers through Streamlit interfaces.

Experience

Work history, roles, and key accomplishments

College AI FAQ Assistant

RAG Application

Built a RAG-based college FAQ assistant that retrieves relevant information from a college knowledge base and generates context-grounded answers using an LLM. Developed document ingestion using LangChain loaders and RecursiveCharacterTextSplitter, generated normalized Hugging Face embeddings, and persisted vectors in ChromaDB.

Python Lang Chain ChromaDB Hugging Face Embeddings GROQ LLaMA Streamlit Data Ingestion Semantic Retrieval

Contract-to-Billing AI Reconciliation Engine

Enterprise RAG & Automation

Architected an automated contract-to-billing reconciliation RAG pipeline that extracts pricing terms, discount clauses, and renewal schedules from complex legal PDFs. Built multi-agent vector retrieval using Qdrant and Supabase pgvector with hybrid search to cross-reference contract terms against active billing JSON payloads.

Python Lang Chain N8n Flowise Qdrant Supabase (PostgreSQL)GROQ LLaMA REST APIs Hybrid Search BM25 Dense Embeddings JSON schema Semantic Similarity

AI Ask My Folder - Multimodal RAG Assistant

Multimodal RAG

Built a multimodal RAG application that searches and answers questions across files stored in user-selected folders. Developed an ingestion pipeline supporting PDF, DOCX, PPTX, CSV, TXT, and image files with semantic chunking and embedding generation.