AI Engineer III (Remote)

Agile Lab is a company founded in 2014 with the mission to create value for its customers in data-intensive environments through customisable solutions that establish performance-driven processes, sustainable architectures and automated platforms based on data governance best practices. Having delivered over 100 successful Elite Data Engineering initiatives, we have used this experience to create Witboost: a modular, technology-agnostic platform that enables modern organisations to discover, value and leverage their data in both traditional environments and fully compliant Data Mesh architectures. With a highly skilled team of over 300 data engineers based in Europe, Agile Lab helps organisations with their data-driven transformation.

Take a look at our handbook to discover our core values and processes.

About the role

We are looking for a Senior AI Engineer who can design, build, and evolve production‑grade agentic systems using agent harness–based architectures .

This role is for engineers who understand that reliable AI agents are not built by prompt engineering alone , but by wrapping capable models in opinionated, lightweight infrastructure : harnesses that provide tools, skills, memory, evaluation, and governance—without over‑engineering orchestration logic.

You will work at the intersection of backend engineering, agent runtime architecture, semantic systems, and modern AI evaluation , building long‑running, self‑correcting agents that operate under real‑world constraints.

RAL: € 48.5 - 62K

Key responsibilities

Build and operate modern AI systems in production

You will help build an AI Operating System : a harness‑centric runtime that turns powerful foundation models into reliable, long‑running systems. This AI OS provides structured capabilities—tool and filesystem access, code‑as‑action execution, skills loaded on demand, semantic grounding via knowledge graphs and ontologies, sub‑agent delegation, memory, and built‑in evaluation loops. Rather than hard‑coding workflows, the OS enables declarative agent design , where behavior emerges from clearly defined capabilities, constraints, and feedback—allowing agents to operate safely under uncertainty, adapt as models evolve, and remain maintainable in production.

Semantic and knowledge-driven intelligence

• Model and leverage semantic layers across data and services:
• Knowledge graphs and entity relationships
• Ontologies and semantic metadata (taxonomy, schemas, reasoning cues)
• Semantic models enabling better retrieval, grounding, and explainability
• Integrate semantic systems with retrieval and agentic workflows to support:
• Grounded answers, better contextualization, traceability
• Consistent domain alignment and reduced hallucinations

Backend engineering excellence

• Build production-grade backend services and APIs that expose AI capabilities
• Design for:
• Scalability, latency, and cost management
• Reliability, observability, and maintainability
• Security and data governance constraints
• Contribute hands-on to implementation, architecture, and engineering standards:
• Clean architecture, modular design, testing strategies
• CI/CD maturity, structured logging, metrics and tracing

Technical judgment and cross-team collaboration

• Evaluate new AI technologies quickly, understanding:
• when to adopt, when to wait, and how to mitigate risks
• Collaborate with product, platform, and data teams to translate requirements into:
• architecture decisions
• delivery plans
• measurable outcomes and quality gates
• Produce high-quality technical documentation and architecture artifacts

Tackle uncertainty and complexity head-on

• Operate effectively in ambiguous problem spaces where:
• requirements are incomplete or evolving
• solutions are not yet clearly defined
• trade-offs must be surfaced early and revisited often
• Break down complex problems into incremental, testable solutions
• Make technical decisions under uncertainty and iterate safely , using:
• experimentation and fast feedback loops
• observability and measurable outcomes
• fallback and rollback strategies
• Help teams and stakeholders navigate complexity with clear technical thinking and pragmatic choices

✅ Requirements (at least 2 must-have)

Strong backend engineering foundation

• Senior‑level experience as a backend software engineer
• Proven ability to build production systems that are observable, evolvable, and maintainable
• Deep understanding of distributed systems trade‑offs

Modern AI & agentic systems

• Hands‑on experience building LLM‑based agentic systems
• Familiarity with:
• agent runtimes
• harness patterns
• long‑running agent behavior
• Clear understanding of pros and cons of new AI capabilities, not hype‑driven adoption

Semantic systems

• Experience or strong familiarity with:
• knowledge graphs
• ontologies
• semantic modeling approaches
• Ability to integrate structured knowledge into agent workflows

Strong productivity in Spring or Python (at least one at pr