I’m a huge fan of Python, TFLite, PyTorch, OpenCV, Streamlit/FastAPI, open-source, and writing testable, reproducible software.
- EduLytics – Low-cost edge-AI classroom engagement (gaze, emotion, drowsiness, yawning) with live overlays & Firebase analytics.
- Traffic Signal Analytics – Signal-phase decisioning + lane/speed analytics served via FastAPI and Docker.
- Spec/RAG Assistants – PDF/spec search using FAISS/Chroma + OpenAI, clean Streamlit UX, and reproducible pipelines.
Languages: Python · C/C++ · JAVA · JavaScript
ML/CV: TensorFlow Lite · PyTorch · OpenCV · MediaPipe
Data/NLP: FAISS · ChromaDB · OpenAI · FireBase
Web/APIs: FastAPI · Streamlit · Flask
DevOps: Docker · GitHub Actions · Poetry
Edge/Hardware: Raspberry Pi · ESP32 · Arduino
- EduLytics / MuACPA – Edge-AI classroom analytics (Pi 5 · TFLite · OpenCV · Firebase)
- Traffic Signal Analytics – Lane/speed detection + phase change API (FastAPI · OpenCV · Docker)
- PipelineIQ / Spec Assistant – PDF & spec RAG search (FAISS/Chroma + OpenAI · Streamlit)
- AI/ML roles (edge-AI, computer vision, applied ML)
- Freelance/consulting for RAG, vision systems, and data products
- Collabs on useful public datasets & open-source tools