sharvesh ak

At Forge Innovation & Ventures, I worked as an Innovation Engineer intern. My recent work includes an AI Visual Quality Inspector that combines YOLOv8 detection with an LLM-based vision model to classify manufacturing defects and assign severity.

I also built a Raspberry Pi 5-based autonomous drone detection system, integrating real-time YOLOv8 inference with Mission Planner flight control. In my deep learning research project, I’m benchmarking EfficientNet, ResNet, and YOLOv8 for diabetic retinopathy severity classification and addressing class imbalance with Focal Loss.

I’m building a multi-agent supply chain risk monitoring system and a RAG-based industrial document Q&A system. As a Mechatronics Engineering student, I bring hands-on experience with hardware alongside AI engineering, from sensor integration to LLM prompt design and vector databases.