Просмотр вакансии

About Us

Silverbee is a fast-moving, seed-stage startup backed by Tier 1 Silicon Valley VCs. After a successful funding round, we are growing our core engineering team to build a production-grade agentic AI platform. Based in Tel Aviv, our mission is to become the service delivery backbone for the next generation of marketing departments.

What you will do

This is a senior individual contributor role with full ownership over production AI behavior, not a research or experimentation-only position.

• Own the end-to-end behavior of a production agentic AI system built with LangGraph, from prompt and architecture design through evaluation, deployment, and production debugging

• Design, implement, and evolve agent architectures (planning, execution, supervision, routing, tool usage) with a clear focus on quality, determinism, latency, and cost

• Take full ownership of system prompt design and tuning, treating prompts as first-class production artifacts that are versioned, evaluated, tested, and deployed as part of the SDLC

• Continuously improve agent performance by running controlled experiments with models, prompting strategies, memory, and tooling, always tied to measurable impact

• Design, run, and evolve evaluation loops (offline evals, regression testing, and production signals) that directly gate releases and drive architectural decisions

• Turn insights from evaluations, traces, and real production usage into concrete improvements in prompts, agents, and system design

• Raise the bar on what “production-ready AI” means at Silverbee: reliability, observability, repeatability, and maintainability

• Mentor other engineers and help establish best practices for building, evaluating, and operating production-grade AI systems

What you need to know

• Strong hands-on experience owning LLM-powered systems that run in production, including diagnosing and fixing real-world failures

• Practical experience designing and operating AI evaluation frameworks, including offline evals, regression testing, and qualitative and quantitative metrics

• Deep understanding of prompt engineering as an engineering discipline: structure, constraints, iteration, debugging, and performance tuning

• Experience with agent frameworks such as LangChain / LangGraph, or equivalent agentic architectures

• Solid software engineering skills (Python and/or TypeScript), including system design, testing, and code quality

• Strong product sense, with the ability to reason clearly about trade-offs between correctness, speed, cost, and flexibility

• Proven ability to operate in ambiguous problem spaces and be accountable for outcomes, not just experimentation

• Clear communication skills and the ability to work effectively with product managers, domain experts, and other non-AI stakeholders

The reason this job vacancy is written in English is that we are looking for a candidate who has proficiency in the language at the B1-B2 level.

If you are interested in participating in our project, please apply for the vacancy! We look forward to meeting you!