cv
Click the button above to download my full CV.
Basics
| Name | Bowen Fang | 
| Label | PhD Student | 
| bf2504@columbia.edu | |
| Phone | 917-293-3109 | 
| Url | https://bwfbowen.github.io/ | 
| Summary | PhD student at Columbia University with a focus on Reinforcement Learning and Autonomous Driving. | 
Work
-  2025.05 - 2025.08 
-  2022.08 - 2022.12 
-  2021.09 - 2022.06 Machine Learning Research Engineer InternAI TOPIAOptimized intraday order placement using ML and RL algorithms.
-  2021.07 - 2021.09 Quant Analyst InternNomura Orient International SecuritiesBuilt FOF platform features, dashboards, and databases.
-  2020.08 - 2020.10 
Education
-  2024 - present Ph.D.Columbia University, New York, NYEngineering and Applied Science- Model Predictive Control
- Diffusion Models AI & RL
- Quantum Computing and Communication
 
-  2022 - 2023 M.S.Columbia University, New York, NYOperations Research- Advanced Big Data and AI
- ML and High-dimensional Analysis
- Robot Learning
- Optimization
- Stochastic Models
 
-  2018.09 - 2022.07 BachelorPeking University, Beijing, ChinaBig Data Management and Applications- Algorithms and Data Structure
- Database
 
Awards
-  2024Calatrava Family FellowshipColumbia University
Skills
| Programming | |
| Python | |
| SQL | |
| C++ | |
| C# | |
| Java | |
| MATLAB | 
| Development | |
| HTML/CSS | |
| JavaScript | |
| Unity | |
| XCode | 
| Machine Learning | |
| JAX | |
| PyTorch | |
| TensorFlow | |
| MuZero | |
| Acme | |
| OpenCV | |
| LLM | 
| Data Science | |
| Apache Spark | |
| Docker | |
| Airflow | |
| MongoDB | |
| Pandas | |
| NumPy | |
| SciPy | 
| Cloud Computing | |
| AWS | |
| GCP | 
| Others | |
| Git | |
| LaTeX | |
| Linux | |
| iOS | 
Languages
| English | |
| Fluent | 
| Mandarin Chinese | |
| Native speaker | 
Projects
-  SINA: Seamless, Intelligent Navigation AnywhereWinner of CS3 VALIDATE Accelerator program, received initial funding and continued funding.
-  Python Open-source Library MuaxImplemented variants of MuZero with JAX/TensorFlow, supported distributed RL and distributed training. Integrated with DeepMind's Acme framework, compared MCTS search policies on Open Spiel multi-agent games (e.g., Go).