working with reinforcement learning, computer vision, and latent representations.
things I'm building:
- dashboard explanation for tree-ensemble models
- deep-learning based ocr for automatic exam digitalization (OBE)
- real-time homography estimation for sports (CV)
- interpretable RL with variational autoencoders and decision trees
- robotics on Raspberry Pi with camera + motor IO
- tucker-based CNN compression
- 📧 email:
ptonso [at] usp [dot] br LinkedIn