Farrel Mahaztra
I've been working remotely as a software engineer for ~6 years while based in Jakarta, Indonesia. Previously, I worked on 10x better spaced repetition software at RemNote.
I've been interested in and concerned about the problem of AI alignment for a long time. I've participated in AI Safety Fundamentals where I worked on sandbagging evaluations and I'm currently working through the ARENA curriculum independently, with the aim of eventually working on technical AI safety research. In terms of research directions, I am particularly interested in model organisms of misalignment and developing more rigorous methods for evals.
During my spare time, I like learning new Japanese and Indonesian recipes, min-maxing strategy games like Civilization and Crusader Kings, reading/watching/thinking about sci-fi, and helping out people interested in starting a career in tech.
I'd like to build and do things that make our world (and future worlds) a better place to be in — less hostile, more creative, more thoughtful, and overall more free.
I'm currently not available for consulting. However, if you'd like to have a chat about anything at all, email me at hello [at] [thisdomain] or DM me on Twitter.

Fusion Gardens of Kepler-186f, Midjourney (2024).