I craft algorithms that help machines see and hear—from real‑time object detection and visual localization to speech enhancement and acoustic scene analysis. My toolbox spans classical DSP and modern deep learning, with a bias for PyTorch‑based research that ships to production.
| Field | Projects | Stack |
|---|---|---|
| Computer Vision | Self‑supervised localization · Lightweight YOLO | PyTorch · MMDetection · ONNX · OpenCV |
| Audio / Speech | Multi‑band speech denoising · Audio‑LLM alignment | TorchAudio · ESPnet · Librosa |
| Edge AI | FPGA/Arm deployment · INT8 quantization | OpenVINO · TensorRT · Vitis AI |
- Languages Python · C++17 · CUDA · Rust (audio‑DSP crates)
- CV Frameworks OpenCV 4 · Kornia · Albumentations
- Audio / Speech Librosa · SoundFile · Kaldi I/O
- Deep Learning PyTorch 2.x · Lightning 2 · Weights & Biases
- MLOps Docker · GitHub Actions · DVC · MLflow
- Visualization TensorBoard · Matplotlib · Streamlit
“In theory, theory and practice are the same. In practice, they aren’t.” — Yogi Berra