I built time-series processing and feature extraction pipelines as an Independent AI/ML & full stack Developer, processing 190,000+ sequential data windows with Python and FastAPI microservices. I also developed mobile telemetry backends with sub-100ms API latency.
At Nokia Solutions and Networks Pvt. Ltd., I trained and quantized YOLO models for PCB fault detection, achieving 93% precision across 10,000+ labeled images. I deployed edge inference pipelines on Raspberry Pi and Linux, streaming anomaly telemetry to a cloud IoT server.
On Hostly, I built a cross-platform mobile application and FastAPI backend for real-time location and attendance verification. My projects also include a time-series scoring platform and a PyTorch Vision Transformer classification pipeline.

