I'm a B.Tech IT student at CEG, Anna University, building at the intersection of AI and hardware constraints. My focus is on running large language models on edge devices where compute and memory are scarce, and designing autonomous agent pipelines that can operate in resource-limited environments. I'm particularly drawn to Defence R&D applications — systems that must be reliable, efficient, and deployable in the field.
- Currently exploring: LLM inference on constrained hardware (quantization, pruning, efficient runtimes)
- Building: multi-agent orchestration pipelines for complex real-world tasks
- Interested in: Defence & aerospace AI, embedded intelligence, RTOS
Languages
AI / ML
Embedded & Hardware
Tools
| Project | What it is | Tech |
|---|---|---|
| nhai-faceauth | Offline on-device face recognition + liveness | React Native TFLite Edge AI |
| roamie | Multi-agent travel planner | smolagents HuggingFace |
| maya | Fully offline voice assistant | Whisper Ollama Kokoro |
| task_master | Self-refining agentic planner | LangGraph Ollama |
| Smart_home_usingLLMs | Natural-language smart home control | LLM Python |
| MotorLeakageFinder_usingML | Real-time engine leak detection | PyTorch Flask |
🔬 LLM inference on edge hardware — quantization, GGUF, ONNX, llama.cpp
🤖 Agentic pipelines — multi-agent orchestration with LangGraph / LlamaIndex
🛡️ Defence AI — reliable, field-deployable intelligent systems
📡 Embedded intelligence — bringing ML to microcontrollers and SBCs
Building at the edge — where compute is scarce and reliability is everything.
Progress [████████░░░░░░░░░░░░] 63/150+
| Category | Solved |
|---|---|
| 💡 DS&A | 57 |
| 💡 Python Coding Interviews | 6 |
🔄 Auto-updated daily · View repo