2026–now
Machine Learning Intern at PathAI
Since August: working on pathology VLMs, distributed post-training, Ray/vLLM inference, and claim-level evaluation.
Multimodal models and agent systems.
I'm Tsung-Huan (Jason), an MSCS student at Georgia Tech, originally from Taiwan, graduating in May 2027. I'm looking for MLE and Research Engineer roles in multimodal models, agent systems, and ML infrastructure.
At PathAI, I work on pathology VLMs, distributed post-training, and inference with Ray and vLLM. Previously at LinkedIn, I built agent context management and its runtime infrastructure. Before that, I worked on VLM pre-training and LLM evaluation at Academia Sinica.
Outside work: workouts, One Piece, and a playlist that can't decide between Chinese pop and R&B.
Selected projects
In progress Taiwan Visual-Language Model 70B
Multimodal data pipelines and pre-training for a 70B VLM, using 50M images and 3B text tokens grounded in Taiwanese culture.
Generative AImultimodallarge-scale pre-training
Research GenAI Safety Assessment
Localized safety evaluation and automated red-teaming for Taiwanese LLMs.
Generative AILLM safetyred-teaming
Course project VisionMealRL
A DQN meal planner for multi-day nutrition goals. I led the RL agent, environment, and baseline evaluation.
Deep RLcomputer visionnutrition AI
Course project Speech Bot
A Jetson Nano robot combining lightweight ASR, OpenAI conversation API, TTS, and hand pose detection for multimodal interaction.
speech processingedge device
Experience
2026
AI/ML Engineer Intern at LinkedIn
May–August: built agent context management and production runtime infrastructure on the Core AI Agent Platform team.
2023–2025
Research Engineer at Academia Sinica
Worked across VLM data and pre-training, LLM safety post-training, evaluator data, and adversarial attacks.
Education
- MSCS at Georgia Tech2025–2027
- B.S. Electrical Engineering, National Taiwan University2018–2022