Public engineering journals, research notes, and lessons learned while building AI systems. Every project in this portfolio has a paper trail.
These articles were written while the systems were being designed, tested, broken, repaired, and deployed. Together they form a chronological record of how the work evolved.
Before building systems, we focused on information quality, memory, uncertainty, and training methodology.
How do you keep AI systems safe after deployment? Open-source security, child safety, and the shield that faces both ways.
Moving beyond model evaluations toward evidence-based AI governance. Seven articles building a complete compliance argument.