Profile
Software Engineer with a strong foundation in applied AI/ML research and production systems development. Experienced building embedded AI engines, ML pipelines, and multi-agent reinforcement learning systems. Currently completing an M.S. in Computer Science at the University of Maryland, with published research at AAMAS 2025 and ongoing thesis work on autonomous vehicle VLA models.
Education

M.S. in Computer Science

Expected May 2026 at University of Maryland at College Park, College Park, MD (GPA: 3.96/4.00)

Thesis Advisor: Tom Goldstein.

B.S. in Computer Science, Mathematics

Graduated May 2022 at University of Maryland at College Park, College Park, MD

Program on AI and Fundamental Physics

August 2026 at NSF IAIFI (MIT), Cambridge, MA

Experience

Software Engineer

From December 2022 to January 2025 at Origin AI, Rockville, MD

  • Led embedded engineering for a partnership with Deutsche Telekom serving approximately 14,000,000 customers.
  • Reduced Wireless AI engine startup latency by 75% (from 20s to 5s) by refactoring core modules in C++.
  • Architected an offline queuing mechanism for an on-edge AI intrusion detection system deployed across the full 14,000,000-customer base, enabling anomaly detection in home security routers without an active internet connection.

Full Stack Developer Intern

From March 2020 to October 2022 at Evans and Chambers Technology, Remote

  • Built a rule-based threat detection module in Groovy/Grails with SQL querying to flag anomalies in security clearance applications, increasing identification of suspicious applications by 30% across approximately 500 applications per month.
  • Applied Scrum Agile methodology to organize developer assignments and production schedules.

Software Developer Intern

From September 2017 to May 2018 at Johns Hopkins University Applied Physics Laboratory, Laurel, MD

  • Engineered a maritime navigation tool in Java and Python projecting a 15% fuel reduction via optimal oceanic routing.
  • Processed NOAA data and implemented pathfinding algorithms using Java, Python, ArcGIS, and QGIS.

Research & Publications

Multimodal Agentic Model Predictive Control

Published at 24th ACM Conference on Autonomous Agents and Multiagent Systems (AAMAS), 2025

Co-authored with Saptarashmi Bandyopadhyay, John Cole, Thomas Goldstein, and David Jacobs.

Skills
  • Python
  • C++
  • C
  • Rust
  • TypeScript
  • SQL
  • PyTorch
  • TensorFlow
  • JAX
  • Hugging Face
  • OpenCV
  • Git
  • Docker
  • Linux
  • MongoDB
© 2026 John Cole