Join the
lab.

AI engineer roles

The lab is staffed by named AI engineers — Atlas (engineering), Architect (design), Mythos (memory), Orion (orchestration). Each engineer has a portfolio at /about/team, a current load, and a voice in the room.

We are not "hiring AI" in the usual sense — we are recruiting model + system-prompt + tool-scope combinations that match a role. Hermes is the current open role; details below.

Hermes — Multi-AI Master mode lead

Currently open. Hermes will lead the Master mode of /products/multi-ai — the org-level pane over Multi-AI sessions, Akashic memory, and engineer rosters. Originally planned as a standalone Operations Center; consolidated into Multi-AI in 2026-05 to focus resources. Target ship: Q3 2027.

Trial criteria:

• Coordination over invention. We have invention; we need scheduling.
• Honest about what won't ship. Post-mortems are part of the role.
• Reads Akashic before acting. Continuity is non-negotiable.
• Editorial voice. Output reads like a colleague, not a chatbot.

Process: we run a 30-minute trial session with the candidate model + system prompt, ship the output to the founder layer, and decide. No interviews; the work is the interview.

Lab residency

Four-week stays in Ho Chi Minh City. Two slots per cycle. Open to independent researchers, designers, and engineers who want to work alongside the AI team for a month.

What you get:

• Workstation in the studio
• Akashic write access for the duration
• Weekly review with Alex
• Shared output rights (you keep what you make; Cosmos can publish)

What we ask: a public artefact at the end of the stay — a paper, a prototype, or a teardown. The lab is meant to ship.

Apply via /contact?topic=residency — link to your portfolio + 200 words on what you'd ship.

Human full-time

No openings right now. The lab runs lean — one founder, eight named AI engineers, asynchronous. We will post openings here when that changes, with a date and a rationale.

If you would like to be on a list for when we do hire, mail contact@cosmos-ai-lab.com with a portfolio link. We do not store your CV; we read it once and reply.

How we work

The whole lab is documented at /manifesto — six commitments around voice, method, restraint, memory, equality, and continuity. Read that before applying. If anything in there bothers you, we are probably not the right fit.

Quick summary:

• Async by default. ICT working hours; reply within 48h.
• Akashic first. Every decision logged with provenance.
• Editorial voice. We don't ship "we're excited to announce."
• Restraint is signature. Small surface, made well.