Agentic AI
CodeAtlas
Multi-Agent Code Intelligence System
A multi-agent system that indexes, understands, and reasons over large codebases using a Graph RAG architecture.
Building intelligent systems powered by LLMs, Agentic AI, Graph RAG, and modern machine learning technologies.
Focused on developing production-ready AI systems, multi-agent workflows, retrieval-augmented applications, and intelligent automation solutions.
01 / About
I'm Abhishek Rajbhar, an AI/ML Engineer passionate about building intelligent systems powered by Large Language Models, Agentic AI, and Graph RAG architectures.
My primary interests lie in LLM applications, multi-agent systems, retrieval-augmented generation, and production-ready AI applications. I enjoy designing intelligent systems that combine retrieval, reasoning, orchestration, and automation to solve real-world problems.
Over the past few years, I have worked on projects involving multi-agent workflows, code intelligence systems, prompt engineering platforms, AI-powered analytics tools, and deep learning applications. My focus is not only on building AI models but also on creating complete end-to-end systems that are scalable, reliable, and deployable.
I was also part of the winning team at Smart India Hackathon 2025, where we developed a solution for a problem statement presented by Bharat Electronics Limited (BEL). Competing against teams from across the country and solving a real-world industry challenge strengthened my interest in engineering practical, impactful technology solutions.
Beyond AI and Machine Learning, I'm particularly interested in the application of intelligent systems to autonomous platforms, drone swarms, and defense technology — exploring how advances in LLMs, multi-agent systems, and autonomous decision-making can contribute to robotics, security, and next-generation defense systems.
B.E. Computer Science (AI & ML)
Theem College of Engineering
University of Mumbai
2022 – 2026
02 / Work
A selection of systems I've designed and built — focused on real, end-to-end AI engineering.
Agentic AI
Multi-Agent Code Intelligence System
A multi-agent system that indexes, understands, and reasons over large codebases using a Graph RAG architecture.
Applied AI
AI-Powered Data Analysis Assistant
Conversational data analyst that performs automated EDA and generates secure, executable analyses from natural language.
LLM Tooling
LLM Prompt Version Control System
Git-style versioning for prompts — track, compare and evaluate prompt iterations across multiple LLM providers.
Deep Learning
Visual Speech Recognition System
End-to-end silent speech recognition using a Conv3D + Bi-LSTM architecture trained with CTC loss.
03 / Toolkit
The stack I use to design, build, and deploy AI systems end-to-end.
04 / Recognition
National-level winner in India's largest innovation competition. Worked on a problem statement presented by Bharat Electronics Limited (BEL).
Active participant in community outreach and volunteer activities.
05 / Contact
I'm always interested in discussing AI, Machine Learning, Agentic AI, Autonomous Systems, and exciting engineering opportunities. Feel free to reach out through email, LinkedIn, or GitHub.
06 / GitHub
Explore my projects, experiments, and AI engineering work on GitHub.