👋 Hi, I'm Pranav
🎓 MS in Data Science · 🤖 Building agentic LLM systems · ⚙️ GPU kernels to production pipelines
I build systems that go from data → model → deployment → real-world use. Currently finishing my MS at UMD (GPA 3.97) while shipping agentic AI systems at MTech Ventures — AutoGen reasoning loops, MCP tool servers, RAG pipelines, and LLM infrastructure for real users.
- Agentic LLM Systems — AutoGen, MCP servers, tool orchestration, reasoning loops
- LLM Fine-tuning & Inference — QLoRA, PEFT, vLLM, multi-GPU deployment
- RAG Pipelines — grounded, production-ready retrieval systems
- CUDA & GPU Systems — kernel optimization, memory hierarchy, edge GPU benchmarking
- Generative Models — diffusion models, segmentation-guided synthesis
- MLOps & Infrastructure — Docker, Kubernetes, ONNX, CI/CD for ML
- Agentic AI systems with real users in the loop
- LLM fine-tuning and distributed inference pipelines
- CUDA kernel benchmarks and GPU systems experiments
- RAG architectures and multimodal LLM apps
- Dockerized, cloud-deployed, production-ready ML systems
- ML / AI · Python · PyTorch · HuggingFace · LangChain · AutoGen · vLLM · PEFT
- Systems · CUDA · C++ · Docker · Kubernetes · AWS · Terraform · GitHub Actions
- Data · SQL · PySpark · Pandas · Pinecone · MLflow
