Shiva Velichalamala

Led large-scale computer vision workflows (OCR, detection, classification) with OpenCV, PyTorch, and TensorFlow. Built end-to-end OCR pipelines using Tesseract, EasyOCR, and PaddleOCR, and developed modular image-processing architectures. Integrated CV outputs into FastAPI microservices, implemented advanced object detection models (YOLOv5, Faster R-CNN), and optimized inference with model pruning and quantization. Established robust evaluation, logging, and production-ready tooling, including Azure DevOps test plans and reusable CV utilities. Enhanced OCR accuracy through hybrid classical and DL feature enhancements, and deployed CV services in containerized, hybrid cloud environments with CI/CD, monitoring, and scalable deployment practices.

Led end-to-end vision-and-language + OCR pipelines for aviation operations, building modular computer vision workflows with OpenCV, PyTorch, and TensorFlow. Implemented multi-stage detection, OCR interfaces, and FastAPI microservices, with automated testing via Azure DevOps and robust logging, optimizing latency for real-time workflow automation.

Delivered a real-time financial risk monitoring platform with modular ML services, including anomaly detection, fraud scoring, and time-series forecasting. Built containerized microservices, centralized model registry, and a feature store; implemented end-to-end CI/CD, governance, and cross-team collaboration with risk and compliance stakeholders.