Cooking AGI
Creator & maintainer ★ 27
Open ML and LLM notes in English and Chinese, with interactive diagrams, coding exercises and interview study plans.
About
I’m a master’s student in computer science at Georgia Tech, graduating in May 2027.
Most recently I worked on LLM-based retrieval at LinkedIn. Before that, at
GT EIC Lab ×
NVIDIA AI Research, I helped build tools for evaluating agents and auditing benchmarks. My research began at
Emory, with directed and time-varying brain networks.
My interests are representation learning, LLM post-training, and model and agent evaluation. I’m particularly interested in how models learn from sparse or noisy feedback.
- Education
Georgia TechM.S. in Computer Science2025–2027
- Emory UniversityB.S. in Data Science2021–2025
Experience
-
LinkedIn
Machine Learning Engineer Intern · LLM-based retrieval for the home feed
-
Georgia Tech EIC Lab × NVIDIA AI Research
Student Researcher · Auditing agent benchmarks ICML 2026
-
Emory · Graph Mining Lab
Student Researcher · Graph learning for brain connectivity AIME, EMBC 2025
Open source
-
— Present
-
—
NVIDIA NeMo RL
Contributor ★ 2.0k
NVIDIA’s framework for LLM post-training and RL; my work is on training efficiency and correctness.
- Cut unneeded compute and communication in log-prob and top-k distillation.
- Worked on multi-reward GDPO in the async training pipeline.
- Worked on streamed PPO minibatches to lower memory use.
Publications
Google ScholarPersonal
Away from work: photography, mountains and rivers, animals, travel and music. I am slowly learning the guitar.
Slide to see the time-lapse