I'm an Agentic AI Engineer who believes the future isn't just "chatting" with AI—it's building systems that perceive, reason, and execute. I spend most of my time bridge the gap between LLM logic and real-world actions.
- Current focus: Moving past basic RAG into high-reliability perception-action loops.
- Philosophy: If it doesn't take an action, it's just a chatbot. I build operators.
- Learning: Vision-Language Models (VLM) for navigating complex UIs.
These are the tools I use to build and break things:
| category | tech |
|---|---|
| agentic | LangGraph CrewAI AutoGen Multi-agent Orchestration Tool-Calling |
| gen ai | Claude 3.5 GPT-4o Advanced RAG Vector DBs Hugging Face |
| infra | Docker Kubernetes AWS GCP CI/CD Linux |
| engineering | Python (Advanced) FastAPI LangSmith PostgreSQL |
Most RAG systems are brittle. I built this via LangGraph to allow for self-correction—if the agent fails to find information, it reasons why and tries a different search strategy instead of hallucinating.
A personal project to solve my own learning curve. Built with Gemini 2.0 Flash, it doesn't just "quiz" you; it syncs with your calendar and creates high-recall sessions based on your actual notes.
Using Pandas and SQL to find the signal in the noise. I processed 1M+ data points to identify 30% growth trends in specific media niches.
- Oracle Cloud 2025 | Generative AI Professional
- Microsoft Azure | AI Engineer Associate
- NVIDIA | Basics to Generative AI
- Google Cloud | Arcade Champion
Whether it's about multi-agent messiness or the future of automation, I'm always down to chat.
- LinkedIn: in/manish2111
- Email: manishchauhan93099@gmail.com
"The best way to predict the future is to build it."