About
I run Golden Toad Research, an independent research group. My work asks whether the numbers used to approve an AI system describe how it behaves in use. A model can score well on overlap, accuracy or a refusal check and still be unreliable on the quantity a clinician or operator depends on. I look for those gaps, measure them and report them as found.
James Irwin
LinkedInPrior research
- Lung Cancer Research Lab, MathematicsTexas Tech, led imaging and ML research
- Cybersecurity Research Lab, Computer ScienceTexas Tech
- EEG research, Electrical EngineeringTexas Tech
Education
- B.S. Computer ScienceTexas Tech, 2026
- B.A. MathematicsTexas Tech, 2026
- Study abroadUVA · Università di Siena
Whether AI evaluation holds up in medical imaging, neuroscience, deepfake detection and cybersecurity.
Active and recent projects, each with a short abstract and a list of goals.
Areas of Interest