About me.
I've been building with AI since 2019. Back then it was neural networks and NLP, before anyone called it agentic. Now I build production LLM apps, stateful agents, and RAG systems with LangChain, LangGraph, Claude, GPT, and Gemini. Same person who ships the backend logic and the frontend you're looking at right now.

Contact
For any sort help / enquiry, shoot a mail and I′ll get back. I swear.
Working Together
I′m usually building something in the agentic AI space. Happy to compare notes or work together if there′s a fit. Resume′s here if you need the formal version. Looking to hire for a project? Get in touch.
Social Links
Hemant Kumar Angajala has been building software professionally since 2020, but his journey into AI started back in 2019 when he got proper training in Machine Learning and Neural Networks at UPX Academy. Since then he's watched the field go from research papers and Kaggle notebooks to something you can actually ship. Agents that hold state. RAG pipelines that don't hallucinate on every third query. Tools people use daily instead of demo once and forget.
These days, Hemant spends most of his time integrating LLMs into real applications. He works with Claude and GPT APIs, builds stateful AI agents using LangChain and LangGraph, and creates conversational interfaces with Rasa and Dialogflow. When he needs to track and debug LLM applications, he turns to Langfuse. For vector search and RAG systems, he's worked with Pinecone, Chroma, and Weaviate. And when projects need custom ML models, he dives into TensorFlow and PyTorch.
On the full-stack side, Hemant is all about Python and FastAPI for backends. It's perfect for integrating AI services. For frontends, he builds with React, Next.js, and TypeScript, focusing on real-time AI interactions and streaming responses. He also builds mobile apps with React Native when projects need to reach users on their phones.
Hemant deploys everything on AWS and Azure with Docker, making sure applications are production-ready with proper error handling, monitoring, and cost optimization. He follows Agile and genuinely likes collaborating with teams. Building intelligent systems is always better with good people around you.
He picks up new frameworks and tools by actually building something small with them, not just reading the docs. That habit started with his AI/ML training at UPX Academy back in 2019, before most of this was mainstream, and it's stuck. He'd rather break something in a side project than in production.
If you need an LLM wired into an existing product, a custom chatbot, semantic search, or a full-stack app where the AI part is more than a marketing bullet point, that's the work Hemant does. Get in touch and describe what you're building. He'll tell you straight if it's a fit.
Backends are Python and FastAPI, built for AI-integrated applications: RESTful APIs with proper auth and rate limiting, database schemas across PostgreSQL, MongoDB, and vector stores, async handling for LLM calls, caching, and microservices deployed with Docker on AWS/Azure — built for reliability and observability, not just a demo that works once.
8 tools
LLMs
Proprietary when it's the right tool, open source when I need control.
7 tools
AI & ML
What I build with most. LLM apps, agents, RAG systems.
5 tools
Vector Databases
Pinecone, Chroma, Weaviate. Whatever the RAG pipeline needs.
8 tools
Frontend
Fast UIs for AI stuff. Streaming responses, no lag.
3 tools
Vibe Coding
Where I actually write code now. Claude Code, every day.
3 tools
Backend
Python, FastAPI. Backbone for every AI service I ship.
8 tools
Others
Git, Docker, Postman. The stuff that holds a project together.
3 tools
Database
Relational or document, whatever the project needs. No preference.
2 tools
Cloud
AWS and Azure. Deployed, monitored, done.
