Senior AI Engineer

Reshark; Amsterdam; On-site; EUR 110,000 - 180,000; Full time

Job summary

Re:Shark builds AI agents for B2B sales teams. We’re hiring our first engineer dedicated to AI to develop the system alongside the CTO.

You’ll work on concrete problems such as entity resolution, context gathering, timing intelligence, and production AI agents. You’ll evaluate models under real constraints, balance quality and cost, and ship with autonomy alongside the CTO.

Qualifications

• Experience building and maintaining LLM-based systems in production.
• Ability to evaluate AI systems using real data, benchmarks, and experiments.
• Understanding of model quality, latency, reliability, and cost considerations.

Responsibilities

• Work with the CTO on the next generation of AI agents for Re:Shark.
• Build crawlers, scrapers, and extraction pipelines to transform messy data into structured context.
• Develop timing intelligence to detect market shifts and drive commercial decisions.
• Design experiments and balance quality, latency, reliability, and cost for AI models.

Skills

LLM-based systems

Backend systems

Data infrastructure

AI evaluation

Tools

PHP (Laravel)

TypeScript

PostgreSQL

Job description

Work with the CTO on the next generation of Re:Shark’s AI agents.

Re:Shark builds AI agents for B2B sales teams. We’re hiring our first engineer dedicated to AI to develop the system alongside the CTO.

What you’ll actually work on

The problems are concrete:

• Entity resolution and context gathering. We maintain a massive graph of companies and people. You will build the crawlers, scrapers, and extraction pipelines to turn messy, conflicting real-world data into structured, trustworthy context.
• Timing intelligence. Companies change constantly. You will build fast, measurable systems that detect market shifts, like new executives or products, and turn that noise into useful commercial decisions.
• Production AI agents. Our agents do not operate in a demo environment; their decisions lead to real outcomes like replies, meetings, bounces, and revenue. You will build the evaluation, tooling, and feedback loops to understand where they succeed and where they fail.
• AI under real constraints. The best model is not automatically the right model. You will evaluate models, design experiments, and balance quality, latency, reliability, and cost, while knowing when an AI model is not actually the right tool for the job.

How we work

• From idea to production. Develop an idea and stay responsible for the solution after launch. Solve the problem wherever it is, even when that means working across several parts of the system.
• The stack. PHP (Laravel), TypeScript, and PostgreSQL. You do not need to arrive as an expert in all of them, but you must be able to jump into unfamiliar codebases and solve the problem.
• Rapid iteration. We build prototypes, test with real data, ship improvements and measure the results.
• High autonomy. You will work directly alongside the CTO, who is deeply hands-on in the codebase. Expect direct communication, fast decisions, and underspecified problems where figuring out the actual issue is part of the job.

You might be a great fit if

• You are an excellent software engineer first, with strong experience building and maintaining LLM-based systems in production.
• You know how to evaluate AI systems using real data, benchmarks, and experiments.
• You understand enough machine learning and statistics to reason about model quality, accuracy, latency, reliability, and cost together.
• You are comfortable spanning backend systems, data infrastructure, and AI.
• You move quickly, take ownership, and prefer shipping and measuring over debating the perfect architecture for three weeks.

This probably isn’t for you if

• You only want to work on model research, write Python, or build AI demos.
• You consider backend, data, or infrastructure to be somebody else’s problem.
• You need detailed tickets before you can start, or you prefer to stay inside one narrow technical specialty.
• You are not comfortable taking responsibility for a system after it ships.