Headline

Generative AI Developer | Shipping Agentic AI with LangGraph, RAG & MCP on AWS | LLMOps · Fine-Tuning · Guardrails | MCA, Presidency College

About

Generative AI Developer with 8+ months of production experience building and shipping LLM-powered applications using LangChain, LangGraph, RAG, MCP Tools, and FastAPI. At Prodevans Innovation Hub, I migrated an HR assistant chatbot (Qwery.ai) from Flask + LangChain to FastAPI + LangGraph, integrating MCP tools with live Zoho People APIs for real-time policy queries, leave management, and employee data retrieval. I designed multi-turn LangGraph state-machine conversations with prompt-level guardrails for safe, policy-compliant responses, and built a separate RAG-based conversational platform using LangChain, Pinecone, and ChromaDB with a Streamlit + REST API frontend and multi-model LLM support. I independently designed, built, and deployed NewsPulse AI end-to-end — a full-stack AI news recommendation system live at newspulse-ai-zs15.onrender.com (backend) and earnest-elf-818b4e.netlify.app (frontend). It runs on FastAPI, MongoDB Atlas, and Sentence-Transformers embeddings, with Groq Llama 3 for automated article tagging, a hybrid recommendation engine (content similarity + collaborative filtering via SVD + recency), JWT/Google OAuth, and GitHub Actions automating news ingestion every 3 hours and daily email digests. I also built FitNova Call Intelligence, an end-to-end sales call analysis pipeline (Groq Whisper + Llama 3.1) that transcribes, diarizes, and scores calls across 5 quality dimensions with automated compliance flagging, backed by PostgreSQL and a role-based Streamlit dashboard — completed as a technical assignment and shortlisted for interview. And VidyaBot, a RAG-based AI math tutor for CBSE students (Class 1-10) built on LangGraph and PGVector, with streaming responses, voice input, and Hindi/Hinglish support. Core stack: Python, LangChain, LangGraph, RAG, AI Agents, MCP Tools, FastAPI, PostgreSQL, MongoDB, PGVector/Pinecone/ChromaDB, Prompt Engineering, JWT/OAuth, GitHub Actions (CI/CD). Pursuing MCA at Presidency College, Berhampur. Microsoft Azure AI Foundry certified. Open to Generative AI Developer / AI Engineer roles — Bangalore,Hyderabad,Remote.

Location

India

AI stack

aistudio

antigravity

claude

gemini

langchain

mcp

Work experience

Generative AI Devloper at Shyadex Technologies LLC, May 2026 - Present · 6 mos, Dover, Delaware, USA 19901, Design and deploy production-ready generative AI systems for a U.S.-based team while working remotely from India. Build agentic workflows using LangGraph to orchestrate multi-step reasoning and tool-calling pipelines. Develop and optimize RAG (Retrieval-Augmented Generation) pipelines for accurate, context-aware AI responses. Implement MCP-based tool integrations to extend LLM capabilities into real business workflows. Deploy and scale AI systems on AWS infrastructure (Lambda, S3, Bedrock) for reliability and performance. Apply prompt engineering and guardrail techniques to improve output consistency and reduce hallucinations. Focus on LLMOps/MLOps practices, including fine-tuning and model distillation, to bridge technical capability with measurable business impact.

Gen Ai intern at Prodevans Technologies, Aug 2025 - Apr 2026 · 9 mos, Bhubaneswar, Odisha, India, I'm a results-driven AI/ML professional specializing in Generative AI and enterprise-grade LLM applications. Currently working as a Generative AI Intern at Prodevance Technology, where I architect and deploy production-ready AI systems that solve real business problems. What I Do: Build intelligent conversational AI platforms using LangChain, LangGraph, and RAG architectures Migrate and scale AI backends with FastAPI for optimal performance Integrate AI agents with enterprise systems (Zoho, PostgreSQL, MongoDB) using MCP tools Design vector database solutions with Pinecone, Chroma DB, and PGVector for semantic search Implement guardrails and prompt engineering for safe, compliant AI interactions Recent Achievements: Migrated HR chatbot backend from Flask to FastAPI, improving scalability and API performance Built RAG-based query platform handling real-time HR workflows with Zoho People API integration Developed VidyaBot, an AI tutor serving NCERT Mathematics curriculum for Class 1-10 students Deployed multi-model LLM systems with context-aware conversational flows using state machines Tech Stack: Python | LangChain | LangGraph | FastAPI | RAG | Vector Databases | PostgreSQL | MongoDB | Azure AI Foundry (Learning) | Transformer Models | NLP Currently pursuing Master of Computer Applications (MCA) from Presidency College, Odisha, and actively upskilling on Microsoft Azure AI Foundry. Open to full-time opportunities in AI/ML Engineering, Generative AI, and LLM Development Let's connect if you're building AI-first products or looking for passionate AI engineers!

Intern at Prodevans Innovation Hub Pvt. Ltd., Aug 2025 - Apr 2026 · 9 mos, Migrated an HR assistant chatbot (Qwery.ai) from Flask + LangChain to FastAPI + LangGraph, integrating MCP tools with Zoho People APIs for real-time policy queries, leave management, and employee data retrieval. Designed multi-turn LangGraph state-machine conversations with prompt engineering and guardrails for safe, policy-compliant AI responses. Built a RAG-based conversational platform using LangChain, Pinecone, and Chroma DB with a Streamlit + REST API frontend and multi-model LLM support. Built NLP and ETL pipelines (NLTK, spaCy, Transformers, Pandas) and worked with PostgreSQL/MongoDB schemas and open-source LLMs via Ollama/Hugging Face.

Education

Master of Computer Applications, Computer Science, Presidency College Berhampur Odisha

Bachelor of Computer Applications, BCA, Presidency College Berhampur Odisha