I'm a barrister (13 years in law, policy, & non-profit boards). 

I've spent 2026 exploring AI safety:
- Currently a SPAR fellow bridging mech interp and law (mentor: Elija Perrier)(https://sparai.org/projects/f26/recLMwVX5XjlH9dx4/)
- started 'Safe AI in Law', an org to build AI safety capability in the legal community (safeai.law). 
- connected with AIS networks in London, Berlin, Cambridge, Lisbon, Phillipines, Australia. 
- Completed the EU Frontier AI & Law Seminar, BlueDot Frontier AI Gov course, and GenStream Residency (I built an applicant screening workflow, rubric, database, PWA for a compute governance course)
- posts on LW & my site (radzaj.com)
- Completed model behavioural research (apartresearch.com/project/behavioural-indicators-of-fault-in-large-language-models-91jb)
- Co-authored an AI incident response paper (apartresearch.com/project/evidence-and-access-in-the-openaihugging-face-incident-wtyv)