Senior AI Engineer, Agents

14 hours ago

Tel-Aviv District, Israel Cyan Full-time


About us
We are an AI Lab that builds agentic systems for real businesses. We work at the cutting edge of the field, turning complex, expert-driven processes into AI systems that run reliably in production. We work closely with our clients to understand their workflows. We design the right architecture for each problem and take it from first prototype to production at scale. We have strong financial backing and a pipeline of real clients and use cases, so we can focus on building things that matter. We are a small, senior team. You will shape the architecture and engineering culture from the early days, and your work will run on real problems from day one.

What you'll do
• Design and build agentic systems and AI architectures for a wide range of business problems, from multi-step workflows to autonomous agents that operate in complex environments like the web.
• Translate messy, real-world processes into well-defined problems: inputs, outputs, success criteria, and where human judgment is still needed.
• Decide, for every component, what should be plain code, what needs an LLM call, what fits a structured workflow, and what really needs an agent. Then build it that way.
• Design tools, skills, state management, and orchestration, including multi-agent setups when they are justified.
• Build human-in-the-loop mechanisms, so systems know when they are confident and when to escalate.
• Own evaluation: benchmarks, golden datasets, LLM-as-judge calibrated against human reviewers, and regression testing for every change.
• Work directly with clients and partners to scope problems and turn them into production systems. What we're looking for
• 5+ years of software engineering experience, with strong production backend skills.
• Hands-on experience building LLM-based systems in production, not only prototypes.
• Practical understanding of agents: tool design, context management, workflows versus agents, and their failure modes.
• Experience with evaluation of LLM systems, or a strong drive to build it properly.
• Clear thinking about trade-offs between cost, latency, and accuracy.
• Comfortable with ambiguity and with owning a problem end to end.
• Docker, Cloud Infrastructure & Services, Grafana, Prometheus
• Browser automation and web understanding (Playwright, DOM, JavaScript-rendered sites). Nice to have
• Crawling or scraping at scale.
• Multimodal models.
• Classic ML background.
• Client-facing or consulting experience.