Benjamin Liu

CS major (AI concentration) at UPenn building and deploying applied AI systems that create meaningful real-world impact. Currently developing agentic models for personalized training, deploying medical ML pipelines with 99% accuracy for department-wide use, and researching transformer-based automation in immunology.

Resume(PDF)

Selected Work (Click For More Information)

A few projects that best represent my work in applied AI and deployed systems.

Aurbit

Startup with agentic AI system deployed to 50+ beta users, generating personalized training insights from Garmin data and co-developed with a Nike professional athlete.

Immunology ML Research

ResNet pipeline achieving 99% accuracy, reducing error detection from hours to seconds, deployed department-wide via Docker and Pennsieve.

ChemoCool

Thermoelectric cooling glove preventing cancer treatment side-effect. Secured a provisional patent, presented to the Stanford Cancer Institute, and recognized as a Conrad Challenge Global Finalist (top 5 in the world).

More Projects

Additional work across research, full stack engineering, and algorithms.

Floya

Deployed productivity platform with Stripe subscriptions and browser-level focus enforcement via a Chrome Extension.

2048 AI

Expectimax agent consistently reaching 4096 and 8192 tiles, with custom Java Swing interface for interactive play and heuristic tuning.

Transformers Research in Medicine

Building transformer-based models for automated cytometric gating.

pace.ai

iOS app published on the App Store generating structured training plans balancing performance gains with injury prevention for distance runners.

Nutrition App

iOS app generating recipes from fridge inventory to reduce food waste and improve nutrition literacy through actionable meal suggestions.

Key Languages, Platforms, and Frameworks Used in Projects

Tap technologies to filter projects. Tap again to deselect.