Apr 2026
In progress
Tetris agent trained with PPO (Stable-Baselines3) on a custom Gymnasium environment. Uses CNN with Rotary Position Embeddings (RoPE), bit-packed uint32 board representation, exhaustive placement search via BFS, and interactive Pygame visualization.
AI
Jupyter
NumPy
Pandas
Pygame
Python
PyTorch
Reinforcement Learning







_169.webp%3Fv%3D1c937692c5d3414d11251d567e3f827140db1475&w=3840&q=75)



.webp&w=3840&q=75)
_169.webp%3Fv%3Dfb85b2b9c4546e514dd553ab04796a3af449012d&w=3840&q=75)
_169.webp%3Fv%3De23714863073cacf25ba2e34c02e370ac5fb8344&w=3840&q=75)

_169.webp%3Fv%3D861e0150dff7ac59c13960c4e812b7de0c13b895&w=3840&q=75)

_169.webp%3Fv%3Df07ebc3cc3f8429151b3b25fb0ee0086915fa53b&w=3840&q=75)

_169.webp%3Fv%3D082664cff25e82037cece14055da802d52dfa639&w=3840&q=75)

_169.webp%3Fv%3D0d299befe0861ec9ef3368b32b29174606fdd300&w=3840&q=75)

_169.webp%3Fv%3D9c4b423a0d8f7ea7b1e63f948705bf35ff03a95a&w=3840&q=75)

_169.webp%3Fv%3D219eafc5d1ddb1fef2b11394c8bb1ed1adb80602&w=3840&q=75)

_169.png%3Fv%3D1641565819dbe591dce2b45420b9ac4c3464eaed&w=3840&q=75)