I'm a Research Manager at SASH, after nine years at Microsoft.
I manage four fellows studying eval awareness and model sabotage, and gave 3 talks on attacks like CoT forgery.
I wanted to learn the business side of things to complement my technical background.
Over three months I explored a market of eval services, met 100+ people in AI safety, and started my own blog.
In one intensive week I saw how monitoring can gain from interpretability through white-box monitoring.
Five days in Berlin on persuasion and power, from Milgram and Asch to practicing how to say no.
I'm a Research Manager at SASH, after nine years at Microsoft.
I manage four fellows studying eval awareness and model sabotage, and gave 3 talks on attacks like CoT forgery.
I wanted to learn the business side of things to complement my technical background.
Over three months I explored a market of eval services, met 100+ people in AI safety, and started my own blog.
In one intensive week I saw how monitoring can gain from interpretability through white-box monitoring.
Five days in Berlin on persuasion and power, from Milgram and Asch to practicing how to say no.
By using this service, you agree to the Terms of Service and Privacy Policy.