I reviewed the annotated output from my vehicle detection and counting project frame by frame against actual traffic, logging missed and double counts to find tracking failures. I fine-tuned Ultralytics YOLOv8s and built a counting pipeline with centroid tracking, direction-aware line crossing, and CSV event logs.
At AgSpert Technologies, I captured drone imagery and labeled 500+ aerial crop images in Roboflow for a disease detection service. I also planned, shot, and edited a traceability campaign video with partner brand Woolah.
At ICAR Central Plantation Crops Research Institute (CPCRI), I set crop-image labeling guidelines and reviewed and corrected labels from a team of 3. I also coordinated field image capture and ran field research with 5,000+ farmers.
I've designed app screens and clickable prototypes in Figma for AgSpeak and a B2B ordering platform, and produced digital content for Vidya AI LLC. My B.E. and M.Tech in engineering, along with hands-on robot building, help me assess whether a physical task actually succeeded.
