Open to opportunities

Chitti User

@kanchanak

Message

AI-ML developer building RAG and agentic retrieval systems with FastAPI.

India

Message

What I'm looking for

I’m looking for a role where I can build production-grade agentic RAG and NLP systems—prioritizing accurate retrieval, fast real-time APIs, and scalable architectures with strong ownership and continuous learning.

I’m an AI-ML developer focused on building retrieval-grounded and agentic AI systems that answer with context, accuracy, and reliability. I design end-to-end, API-driven architectures and bring models into production using modern backends and deployment pipelines.

In my current role, I’m developing a hybrid RAG QA system that combines sparse (BM25) and semantic retrieval, with web fallback for broader coverage. I integrate OpenAI-compatible LLMs for controlled responses and implement full frontend-to-FastAPI backend flows, including responsive chat experiences and user configuration dashboards.

Previously, I built an agentic university Q&A chatbot using FastAPI, LangChain, MongoDB, and local LLMs—using hybrid retrieval, tool-calling agents, and real-time multi-turn responses. I also developed an end-to-end disaster tweet analyzer with NLP/ML, including SMOTE for class imbalance, and delivered a web interface with geospatial visualization.

Experience

Work history, roles, and key accomplishments

Current

AI-ML Developer

Current

Haystek Technologies

Oct 2025 - Present (7 months)

Building a hybrid RAG-based QA system combining BM25 and semantic retrieval with DuckDuckGo web fallback. Developed an end-to-end API-driven architecture with a FastAPI backend and deployment pipeline to power a responsive configuration and chat interface.

Hybrid RAG BM25 DuckDuckGo API fastAPI Retrieval Augmented QA

AI-ML Intern

Haystek Technologies

Jul 2025 - Sep 2025 (2 months)

Developed an agentic RAG chatbot for academic question answering for students, parents, and faculty. Implemented hybrid dense+sparse retrieval with tool-calling agents, web data scraping/structuring, and a low-latency FastAPI + MongoDB backend.

Agentic RAG fastAPI MongoDB Dense Retrieval Web Scraping

AI Intern

Infosys Springboard

Oct 2024 - Dec 2024 (2 months)

Built an end-to-end disaster tweet analyzer to identify, classify, and analyze disaster-related tweets using NLP and ML. Created synthetic datasets and addressed class imbalance with SMOTE, then deployed a web interface for real-time analysis and geospatial visualization.

NLP Machine Learning Disaster Tweet Classification Synthetic Data Generation SMOTE Python