Job description:

Design and operate AI agents that reason, act, and collaborate with humans

Build agent orchestration systems (tool use, memory, planning, feedback loops)

Run large-scale inference and training pipelines in production

Work on model evaluation, drift detection, and continuous improvement

Optimize for latency, cost, and reliability at real enterprise scale

Partner closely with DevOps, security, and backend engineers

Ship models that are auditable, explainable, and safe in sensitive environments

Hard requirements (non-negotiable):

5+ years in ML / AI / Applied Research roles

Hands-on experience fine-tuning large models (LLMs or multimodal)

Deep familiarity with agent architectures (tool use, memory, planning, reflection)

Real production experience

Heavy, daily usage of AI coding tools (Claude, Codex, Cursor, etc. — this is how we work)

Experience operating models at scale (high throughput, real traffic)

Strong signals you're a fit:

You've shipped agent systems that run unattended in production

You've fine-tuned models for precision, not just demos

You think about evaluation frameworks as much as training

You care about failure modes, hallucinations, and abuse cases

You prefer fast impact over papers

Comfortable working 5 days a week from our Tel Aviv office

Jobs