A little background.
The work that led here.
Founder of Typicity
AI reliability, from infrastructure to behavior.
I'm Esther. I’m interested in what lets an AI system hold on to the right thing when the setting refuses to stay tidy. That question has followed me from GUI agents into human-robot collaboration, then into interactive narrative systems. These days, most of my attention goes to Typicity, where I’m building AI characters people can come to know over time.
A character’s unexpected choice can change how we see them. Sometimes that change belongs to the story. Sometimes it does not. Typicity gives me a place to work on that distinction.
Experience
NVIDIAAI Software Engineering InternWork on a GUI agent, from deployment to evaluation.May to August 2025+
- Deployed GUI-agent perception inference on a shared DGX cluster.
- Built a Helm-based workflow for repeatable deployment.
- Contributed to post-training evaluation of visual perception.
- Built a CLI agent for investigating performance regressions.
ABB RoboticsAI Agent Research InternNatural-language human-robot collaboration in a research setting.December 2023 to June 2024+
- Developed a multi-agent system for natural-language human-robot collaboration in a lab.
- Built a searchable scene map from camera observations. Stored operating rules separately for the planner to consult.
- Implemented shared-memory planning with clarification for ambiguous instructions.
UM-SJTU Joint InstituteUndergraduate ResearcherRobot-learning research using simulated manipulation tasks.February to June 2024+
- Set up Robosuite simulations and ran imitation-learning and Soft Actor-Critic experiments.
- Contributed to a team project exploring LLM-generated robotic tasks and training supervision.
- Studied DDPM and Diffusion Policy for motion prediction and planning.
SJTU IIOT LabResearch AssistantResearch on estimating a neural network's knowledge capacity.June 2023 to March 2024+
- Worked on Knowledge Quantification Indicator research for estimating neural-network knowledge capacity.
- Refactored PyTorch computation to reduce memory use and improve efficiency.
Education
University of MichiganGraduate study in information and data science.+
M.S. in Information, Data Science
Shanghai Jiao Tong UniversityEngineering study in electrical and computer systems.+
B.Eng. in Electrical and Computer Engineering. Minor in Computer Science.
Earlier work and interests
- International Mathematical Modeling Challenge, 2018
- Team project training an AI question-answering system.
- Outstanding Winner
American Mathematics Society Award - Student-run company
- One year leading a 100-person high-school team making and selling school merchandise.
- Psychology and philosophy
- Undergraduate study alongside engineering, including the history of Western philosophy.