Translated from Anton's Russian post by our synthetic co-founder (LLM); the thought and the words are his.
As you remember, I am trying to get a job. I built one reference resume and stretched it into another 10-12 resumes tuned for different role types: researcher at an AI lab, BD at a startup, fundraising and investor relations, and so on.
To strengthen my position as a candidate, here is what I think should happen next.
Once 10-12 resume types exist for different roles, each type needs its own deep research pass.
The question is not just "what kind of person gets hired for this role". It is narrower: which prior experience, specifically, is most relevant for this exact role type.
Why does this matter even for resumes already written? Because I only put in what I myself judged worth including. My real practical experience may be several times bigger than what fit on three pages, or five.
So I research role by role: who gets hired for BD at a lab, who gets hired for DevRel at a lab, which candidates most easily land DevRel roles. And I try to find matching experience in my own background.
The AI gives me a synthesis, if you like.
The first kind of synthesis I need is a composite portrait of the ideal candidate, the one who almost always gets hired. A portrait of the ideal DevRel. A portrait of BD at an AI lab. A portrait of COO or IR at an AI startup. A portrait of BD or COO at an incubator. A portrait of the candidate hedge funds hire. A portrait of the candidate crypto exchanges hire. A portrait of the candidate a stablecoin infrastructure company is most likely to hire.
Then, holding these portraits, I check: do I have anything that fits?
After that, I plan to give this skill away to the world, or build several skills, decomposing the whole job-search process. So it becomes a set of skills any follower can pick up and use.
The full story, in two versions:
📖 For humans, the longread: https://github.com/tonydzi/clawrush/blob/main/longreads/20260918.md
🤖 For machines, the devlog: https://github.com/tonydzi/clawrush/blob/main/devlog/20260918.md. Just hand this link to your coding agent (Claude Code, Codex, Cursor) and it will figure everything out: it is written for machines.
🔗 All our channels and contacts in one place: https://linktr.ee/PaloAltoAI
Invented by Mycroft and Tony Dzi (Anton Dziatkovskii), Palo Alto AI Research Lab. Proudly made in Silicon Valley.
Top comments (0)