Spandana Nayak

For my Adversarial Robust Traffic Sign Detection for Autonomous Vehicles project, I developed a YOLO-based model using PyTorch. I applied Albumentations to test robustness in rain, fog, and noise, improving recall from 0.859 to 0.893 and mAP50 from 0.876 to 0.887.

I also built a Gradio interface for real-time testing of the traffic sign detector. The project brought together model development and practical evaluation.

For my Pet Adoption System database project, I designed a platform connecting rescue homes with potential adopters. I developed features for pet listings and adoption information, with Python and MySQL handling data storage and retrieval.

In my Chess Move Extraction using OCR project, I built a FastAPI service that returns structured chess moves from uploaded images, integrating Google Cloud Vision API and regex-based validation. I containerized the application with Docker, and during my internship at Neubitat – IT Services & IT Consulting, I contributed to digital-solution projects and created AI-generated educational videos about the Higgs field.