Rachel Dong
Staff Research Engineer at Iconic
San Francisco, California, United States
Summary
🤩
Rockstar🎓
Top SchoolRachel Dong is a Staff Research Engineer in San Francisco with nine years of experience building and shipping ML and generative AI systems for games and entertainment. She blends research-grade expertise in LLMs, generative models, 3D computer vision, and reinforcement learning with production engineering—work that spans Unity Metacast volumetric capture, game-focused generative agents at Riot, and scene understanding for autonomous systems. A significant open-source contributor to Unity’s ML-Agents, she expanded visual encoders (Nature CNN, ResNet) to improve training with visual observations, demonstrating both systems-level thinking and hands-on model work. Rachel holds an MS from Carnegie Mellon and brings a rare combination of academic rigor and product-focused delivery across studios and startups.
9 years of coding experience
7 years of employment as a software developer
Master of Science - MS Computer Science, Master of Science - MS Computer Science at Carnegie Mellon UniversityBachelor's degree Electrical Engineering Technologies/Technicians, Bachelor's degree Electrical Engineering Technologies/Technicians at National Tsing Hua University
English, Chinese
Github Skills (10)
neural-network10
mask-rcnn10
faster-rcnn10
deep-reinforcement-learning10
machine-learning10
deep-learning10
tensorflow10
python10
unity3d10
resnet9
Programming languages (2)
Github contributions (5)
Unity-Technologies/ml-agents
Jul 2019 - Sep 2021
The Unity Machine Learning Agents Toolkit (ML-Agents) is an open-source project that enables games and simulations to serve as environments for training intelligent agents using deep reinforcement learning and imitation learning.
Role in this project:
ML Engineer
ML EngineerContributions:2 releases, 275 reviews, 417 commits in 2 years 2 months
Contributions summary:Rachel's commits focused on enhancing the Unity Machine Learning Agents Toolkit (ML-Agents). Their work involved adding different types of visual encoders, specifically Nature CNN and ResNet, to the training models. The changes included modifications to the Python code, including the `models.py` and `ppo/models.py` files, to support these new encoder types and refactoring the `vis_encoder_type`. This significantly expanded the toolkit's capabilities for training agents with visual observations.
servedeep-reinforcement-learningtrainingunityreinforcement
NTHU-EE-CV-2016-Fall/homework0
Sep 2016 - Jul 2019
Contributions:2 PRs in 2 years 10 months
Find and Hire Top DevelopersWe’ve analyzed the programming source code of over 60 million software developers on GitHub and scored them by 50,000 skills. Sign-up on Prog,AI to search for software developers.