Tharun kumar

I’m an early-career AI engineer with hands-on experience shipping production workflows that connect AI services to reliable APIs and data pipelines. I focus on building LLM-backed features with clear contracts, iterative validation, and dependable runtime behavior.

In my student internship, I built and maintained REST API integrations between frontend components and backend AI services using React and Tailwind CSS. I also supported LLM-backed feature delivery pipelines—aligning frontend data contracts with evolving backend AI interfaces while collaborating in agile sprint ceremonies.

I’ve designed end-to-end AI systems and production-grade pipelines, from multi-stage agentic inference to ML deployment. Highlights include DAMAGESCOPE (multi-modal Siamese ResNet-18 + SAM ViT-B fusion with IEEE RMKMATE acceptance) and Scrape-Tober (a FastAPI + Docker pipeline with orchestration, monitoring, retries, and DevSecOps-oriented validation).

My work reflects an engineering ethos: ship systems that are not only accurate, but observable, modular, and robust under real constraints. I enjoy taking models from prototype to hardened delivery—whether it’s TFLite deployment on embedded hardware or reliable agentic workflows with state tracking.