Hi! I'm Bohan Yang, a junior at Cornell University, studying Computer Science with minors in Artificial Intelligence and Business.
I'm passionate about leveraging AI to make the world a better place.
Currently, I'm a robotics researcher at the Cornell EmPRISE Lab, advised by Prof. Tapomayukh Bhattacharjee. My recent work includes developing an Adaptive Robot Query System for Task Failure Recovery, which enables robots to proactively ask for human help when uncertain, and autonomously learn recovery strategies from failed tasks.
Previously, I researched on LLM-Guided Differential Equation Inference from Cornell MAE department and Machine Learning for Concussion Diagnosis from Harvey Mudd College.
In the summer of 2024, I co-founded Pixelate Inc., a startup focused on developing a CAD Copilot for engineers, designers, and makers. Please check out our website to learn more about our work!
I'm a researcher, an entrepreneur, and a dreamer. If you share the same passion, feel free to reach out!
Computer Science @ College of Engineering
Cornell University
Timeline
Pixelate
2024 - Present
Co-founder developing AI-powered CAD Copilot for engineers, designers, and makers
Cornell EmPRISE Lab
Summer 2023 - Present
Robotics Researcher working on Adaptive Robot Query System for Task Failure Recovery
Dasion
Summer 2023 - Present
Machine Learning Software Engineer developing intelligent voice agents for elder care
Cornell MAE Department
2023
Research Assistant working on LLM-Guided Differential Equation Inference via Symbolic Regression
Cornell University
2022
Accepted into the College of Engineering at Cornell University.
Harvey Mudd College
2021 - 2023
Machine Learning for Concussion Diagnosis from Electroencephalogram (EEG) data.
Publications
Journal Papers
-
Gu, W., Yang, B., & Chang, R. (2023). Machine Learning-based EEG Applications and its Applications. Journal of Neurology & Neuroscience, Vol. 14 No.S6:002.
Preprints
-
Gu, W., Chang, R., & Yang, B. (2022). EEG Machine Learning for Analysis of Mild Traumatic Brain Injury: A survey. arXiv preprint arXiv:2208.08894.
Projects
DXF2CAD
A generative AI framework that learns from 2D DXF blueprints to construct editable 3D CAD models in STEP format in seconds.
Technologies: Python, PyTorch, Transformers
CADParser
Reimplementation of a transformer-based model that generate parametric construction sequences from 3D CAD models.
Technologies: Python, PyTorch, Transformers, Graph Neural Networks (DGL)
Parallelizing Diffusion Modeling with MPI
A high-performance parallel simulation of 2D diffusion using MPI.
Technologies: C++, MPI, High Performance Computing
Address Matching
An address matching system to ensure the consistency of healthcare provider directories.
Technologies: Python, Pandas, Locality Sensitive Hashing (LSH)
Pokémon Kingdom
An interactive turn-based Pokémon battle game.
Technologies: OCaml