About Me
Hey, I’m stillonearth, previously cwiz and still cwiz elsewhere. I’ve been on a startup journey, a corporate path and now diving into exciting worlds of AI, Reinforcement Learning, Robotics, Simulators, Game Engines.
Open-Source Libraries
bevy_rl
Build Reinforcement Learning Gym environments with Bevy to train AI agents that learn from screen pixels
https://github.com/stillonearth/bevy_rl
bevy_mujoco
Build robotic simulations with Bevy and MuJoCo
https://github.com/stillonearth/bevy_mujoco
veilid_duplex
Full-duplex asyncronous communication between two peers with Veilid
https://gitlab.com/cwiz/veilid_duplex
bevy_veilid
Build two-player turn-based p2p games with Bevy and Veilid
https://github.com/stillonearth/bevy_veilid
Research Projects and AI Recipes
Continous Control with Deep Reinforcement Learning for Unitree A1 Quadruped Robot
2022, October
https://github.com/stillonearth/continuous_control-unitree-a1
Multi-Agent FPS Gym Environment with bevy_rl
Bevy Jam 1 Entry
2022
https://github.com/stillonearth/bevy_rl_shooter
Checkers Game on Rust and Bevy, OpenAI Gym Environment and AI Agent based on Monte Carlo Tree Search (MCTS) with neural heuristics (AlphaZero) on Python and PyTorch
2022, January
https://github.com/stillonearth/CheckersOnBevy
Reimplementing Alpha-Zero for board game of Go
2019, December
https://github.com/cwiz/guided_monte_carlo_tree-search/blob/master/Tree-Search.ipynb
Using Policy Optimization (PO) Algorithms for Solving Zero-Sum Games with Self-Play
2019, September
https://github.com/cwiz/DRLND-Project-Collab-Compet/blob/master/WRITEUP.md
Using Policy Optimization (PO) Algorithms Soft Actor-Critic (SAC) and Proximal Policy Optimization (PPO) for Solving Unity ML Reacher and Crawler Continous Control Environments
2019, August
https://github.com/cwiz/DRLND-Project-Continuous_Control/blob/master/WRITEUP.md
Using Deep-Q-Networks (DQN) for solving Unity ML Agents Banana Collector Discreet Control Environment and Evaluating DQN Improvements
2019, July
https://github.com/cwiz/DRLND-Project-Navigation/blob/master/WRITEUP.md
Semantic Segmentation with Fully-Convolutional Neural Network for Identifying and Following a Human Target by a Quadcopter
2019, May
https://github.com/cwiz/RoboND-DeepLearning-Project
3D Perception with Point-Cloud Camera for Willow Garage PR2 Robot in Gazebo Virtual Environment and Robot Operation System (ROS)
2019, March
https://github.com/cwiz/RoboND-Perception-Project/blob/master/README.md
Kinematics Pick & Place for KUKA KR210 Robotic Arm in Gazebo Virtual Environment and Robot Operation System (ROS)
2019, March
https://github.com/cwiz/RoboND-Kinematics-Project/blob/master/README.md
Mars Rover Perception and Decision Making for Solving Search and Sample Return
2018, December
https://github.com/cwiz/RoboND-Rover-Project/blob/master/README.md
Quadcopter State Estimation with Kalman Filtering in C++
2018, October
https://github.com/cwiz/udacity-flying-car-cpp-estimation/blob/master/WRITEUP.md
Quadcopter Motion Planning in Urban Environment
2018, September
https://github.com/cwiz/udacity-flying-car-motion-planning/blob/master/WRITEUP.md
PID Controls for Quadcopter in C++
2018, September
https://github.com/cwiz/udacity-flying-car-cpp-controls/blob/master/WRITEUP.md
NN Models & Experiments
AI-Driven Pirate Video Translation of YouTube Videos 90s VHS Style
2023, August
https://github.com/PiratesAye/pirate-neural-translate
jax-rl
Compact one-file implementations of some RL algorithms in JAX
https://github.com/stillonearth/jax-rl
Games
Legend of Mierda
Game to support Russian-Mexican polytical prisoner AlexTime
2024, March
https://cwiz.itch.io/legend-of-mierda
Education
Moscow Aviation Institute
Engineer
2006 — 2013
Udacity
2018 — 2019
Classes:
- Flying Car and Autonomous Flight
- Robotics Software
- Deep Reinforcement Learning
- AI for Trading
Employment
self-employed
Developer
2015 — now
Neural networks, reinforcement learning and robotics
Indoor Navigation and footfall analytics with wireless singal probing
Online hotel booking service