Yash Gaikwad

I’m an AI/ML Engineer with 3+ years of hands-on experience shipping production-grade systems across Generative AI, LLMs, RAG pipelines, computer vision, and edge AI. I focus on turning research-quality ideas into reliable, deployable products.

I design LLM-powered applications using Azure OpenAI and Ollama, leveraging prompt engineering and agentic AI architectures to improve coherence and reduce hallucinations. I also build context-augmented solutions with vector databases like FAISS and ChromaDB to make answers grounded and relevant to domain queries.

On the computer vision side, I’ve built end-to-end pipelines from detection to OCR—using YOLO and PaddleOCR—and optimized them for faster inference with ONNX, TensorRT, and OpenVINO. I’ve deployed real-time CCTV/embedded solutions under strict resource constraints, including sub-128MB deployments.

My professional ethos is practical engineering: measure performance, optimize for latency and memory, and deliver systems that actually work in production. I enjoy building end-to-end experiences, from model pipeline and evaluation to integration with backend services.