SlipstreamJobs tracks this role from the company's public career site. Apply directly on the employer's site.
ZenML is an open-source framework for production ML and agentic pipelines. Kitaru is their new product that turns agent traces into reliable evaluations—ingest production sessions, mine high-signal cohorts, mock tools, replay agents, and ship with confidence.
You will build product for agent builders on Kitaru, focusing on evals, replay, cohorts, and developer experience. You'll ship features with taste and judgment, knowing when something is sharp versus theater. You'll create in public on X (Twitter), where the AI engineering audience lives—posting, replying, demoing, and pulling signal back into the product. Your GitHub/GitLab history is part of how the company found you.
You'll wear different hats week to week: marketing motion, feature work, customer demos—whatever moves the needle. You're not siloed as "the marketing person" or "the eng person." You're the person who makes progress. You'll stay close to the ICP: AI agent builders, eval nerds, people wiring LangGraph, Claude Code, custom runtimes. You already hang out there; this job pays you to go deeper.
You'll feed the loop: what you hear on X, in demos, and in issues becomes product. What you ship becomes the next thread. No wall between build and distribution.
Success at 6 months means: a visible, credible presence in the AI engineering / agent-builder conversation (people associate you and Kitaru with useful takes and real demos, not noise); features you owned that agent builders actually use (shipped, not slideware); demo and content motion that creates pipeline and product signal without a separate GTM person holding your hand every week; and a raised bar for what the team ships and says in public because you were in the room.
You are an AI engineer who builds in public—not a pure marketer learning to code, not a pure engineer who hates posting. Your X and GitHub/GitLab are your résumé. You have high agency and low ego; you'll invent the week's job. You're technical enough to ship, comfortable in Python and agent frameworks, evals, traces, and tooling. You can demo live and debug with a customer without hiding behind slides. You have taste for builders and know the difference between a viral thread and a useful one. You'd rather show a sharp demo than a brand campaign. You're already in the water, following and shipping near people building agents. You learn by default because the field is changing monthly—new models, new harnesses, new ways to run and evaluate agents. You keep up because you can't help it, try things before they're documented, and bring what you learn back to the team and into the product.
The company offers a direct line to founders, real ownership on a new product, stock options, competitive compensation, and a team that already ships OSS used by thousands of engineering teams.
REQUIREMENTS:
- AI engineer who builds in public (X and GitHub/GitLab presence required)
- Technical proficiency: Python, agent frameworks, evals, traces, tooling
- Ability to demo live and debug with customers
- Understanding of AI agent builders, eval practices, and production agent deployment
- High agency, low ego, ability to invent and prioritize work independently
- Already embedded in the AI engineering / agent-builder community
- Nice to have: OSS contributions people actually use; production agent building experience (not just tutorials); prior startup experience; design/DX instincts; documentation and API design sense
