I build practical AI systems with retrieval, agents, forecasting, APIs, and clean product interfaces.
- Integrated M.Tech CSE student at VIT Vellore
- Currently working on demand forecasting, multimodal RAG, and agentic AI workflows
- I like building AI systems that move from experiments into usable software
- Most comfortable across Python, FastAPI, React, LangGraph, RAG, and backend systems
- Interested in applied AI, developer tools, automation, and products that make complex workflows simpler
focus:
- demand forecasting for retail planning and operational signals
- multimodal RAG pipelines with voice, text, retrieval, reranking, and caching
- multi-agent workflows with tool routing, approvals, and real-time dashboards
- full-stack AI products with clean APIs and usable interfaces
learning:
- better LLM evaluation patterns
- scalable retrieval architectures
- production observability for AI systems
| Area | Tools |
|---|---|
| AI / ML | LLMs, RAG, LangGraph, LangChain, TensorFlow, PyTorch, OpenCV, YOLO |
| Data | PySpark, SQL, feature engineering, forecasting, evaluation |
| Backend | Python, FastAPI, Flask, PostgreSQL, Redis, REST APIs |
| Frontend | React, TypeScript, JavaScript, Tailwind CSS |
| Infra | Docker, AWS,Azure, Git |
| Project | What it is | Stack / Ideas |
|---|---|---|
| Astra | Multi-agent personal assistant platform for workflow automation | LangGraph, FastAPI, React, real-time execution |
| Multimodal RAG Pipeline | Voice + text retrieval system with generated answers | Hybrid search, reranking, caching, modular ingestion |
| Fridge Vision | Ingredient detection to recipe recommendation workflow | YOLO, computer vision, recommendation logic |
| Scalable RecSys Engine | Recommendation system designed for large datasets | Similarity search, KMeans, caching, 25M+ record design target |
Portfolio · GitHub · LinkedIn · Email
Building AI systems that actually ship.