At Quadrant Technologies, I contribute to real-time, client-facing GenAI applications, integrating RAG pipelines with LLMs and Transformer-based architectures. I also develop model-backed functionality through REST APIs and FastAPI services for real-time inference.
I built and tested an AI-driven skin analysis prototype that detects and grades skin conditions from live video or images. It combines OpenCV face detection, SSD-based lesion localization, handcrafted features, and CNN classifiers implemented in PyTorch and TensorFlow.
For Signalrank, I built a deterministic candidate-ranking engine that scored more than 100,000 profiles within a five-minute runtime and a 16GB RAM, CPU-only budget. I’ve also co-built a voice-first AI interviewer with persistent interview state and session recovery.
