中文
ArticleTalkView sourceHistory

University of Pennsylvaniaedit

Graduate study of Colar Wang (2025–2027)

This article documents Colar Wang's graduate study at the University of Pennsylvania (UPenn), which he entered in August 2025 as a candidate for the Master of Science in Systems Engineering.1

Programedit

The Master of Science in Systems Engineering is offered by UPenn's Department of Electrical and Systems Engineering, within the School of Engineering and Applied Science. The program sits at the intersection of applied mathematics, machine learning, and systems modeling — a combination Wang has described as "engineering grammar for people who build AI products."

Courseworkedit

Wang's coursework, drawn from the systems-engineering and data-science tracks, has emphasized:

  • Applied Machine Learning
  • Statistics for Data Science
  • Simulation Modeling
  • Marketing Analytics

Project leadershipedit

Campus-Scale Img2GPS Localization (2025)

As Project Lead between October and December 2025, during his first semester at UPenn, Wang directed the development of a Campus-Scale Img2GPS Localization system — a high-precision image-to-GPS geolocation model intended for campus-scale navigation. The project reported a 29.7 percent reduction in localization error relative to its baseline, and achieved a 2.5-fold acceleration in model convergence through a rigorous trade-off analysis between computational overhead and spatial accuracy.

A defining technical contribution was a stage-wise adaptation strategy that reduced the model's parameter count by approximately 96 percent, enabling deployment in resource-constrained environments without compromising core accuracy. Wang has cited this compression experience as directly relevant to the edge-side design decisions later adopted in KitchenSurvivor. The full project report is available as a PDF.

Context within the broader biographyedit

Wang's move to Philadelphia in August 2025 immediately preceded the founding of KitchenSurvivor in November 2025, and coincided with the conception of AgentConfig in early 2026. Both products draw methodologically on the quantitative and modeling foundations of the program, and Wang has cited the simulation-modeling and applied-ML coursework — as well as the Img2GPS project above — as directly relevant to the edge–cloud inference pipeline he designed.

See alsoedit

Referencesedit

Footnotesedit

  1. "Master of Science in Systems Engineering". University of Pennsylvania — Electrical and Systems Engineering. ↩