AI Tech AI Lead Hoofddorp · hybrid

Devleaps; Hoofddorp; Hybrid; EUR 120,000 - 160,000; Full time

Job summary

Devleaps is seeking an AI Tech Lead to set the architecture for reliable AI capabilities embedded in client systems. You will define model serving, retrieval, orchestration, and platform services, while enforcing guardrails and monitoring.

You will work with client teams to elevate production AI features and ensure robust, scalable solutions. You will drive evaluation pipelines, cost modeling, and latency budgets with the rigor of a production system, designing for graceful degradation and clear

Qualifications

• 6–10+ years of software engineering with lead/principal experience.

Responsibilities

• Lead technical direction on AI engagements: architecture of AI estate and platform services.
• Build and enforce guardrails: input/output filtering, content policies, audit logging, human-in-the-loop checks.
• Embed as a technical anchor in client teams: pair with engineers on production AI features and review work to a common standard.
• Own evaluation and observability as architecture: pipelines, drift detection, cost per inference, latency budgets.
• Design for failure: fallback behaviour, rate limiting, circuit breaking for AI features.
• Translate AI capabilities into CTO-facing decisions, clarifying what is possible with current data.
• Identify bottlenecks and remove work that slows teams down.

Skills

Architectural leadership

Lead engineering

LLM/ML systems

Python

AI tooling

Cloud platforms (Azure)

MLOps

API design

Stakeholder management

Tools

LangChain

LangGraph

LlamaIndex

OpenAI APIs

Anthropic

Azure OpenAI

Bedrock

Job description

Engineering is human. So are you. AI just makes the loop faster.

At Devleaps, we help clients go from 'we tried a chatbot' to 'we have a reliable AI capability embedded in our systems'. We're looking for an AI Tech Lead who sets the architecture for that capability, sets the guardrails for how it is built and used, and makes sure both hold up under load.

WHAT YOU'LL DO

• Lead technical direction on AI engagements: own the architecture of the AI estate - model serving, retrieval, agentic orchestration, and the platform services underneath - and set the engineering standards it is built to.
• Build and enforce guardrails: input and output filtering, content policies, audit logging, human-in-the-loop checkpoints, and review conventions for AI-generated code.
• Embed as a technical anchor in client teams: pair with engineers building their first production AI feature, and review both human and model-authored work to the same bar.
• Own evaluation and observability as architecture, not afterthought: evaluation pipelines, drift detection, cost per inference, and latency budgets treated with the rigour of any other production system.
• Design for failure: fallback behaviour, rate limiting, and circuit breaking for AI-dependent features, so a bad model response degrades gracefully rather than breaking the product.
• Own the technical narrative: translate what AI can and cannot reliably do in this stack into decisions a CTO can sign off on - including when the honest answer is 'not with this data'.
• Remove what slows teams down: identify systemic bottlenecks, including AI workflows that promise speed but quietly add rework.

YOU ARE

• A seasoned engineer who has led: you've made architectural calls, mentored engineers, and owned outcomes - not just delivered tickets.
• A consultant at heart: you ask before you prescribe, push back with evidence, and are as comfortable saying an LLM is the wrong tool here as you are proposing one.
• Comfortable with ambiguity: legacy constraints, shifting priorities, and a fast-moving AI tooling landscape do not throw you off course.
• A multiplier: you measure success by how much better the team and the codebase get, not just by what you personally shipped.
• Human, with opinions and a healthy scepticism about AI hype.

YOU'LL BRING

• 6-10+ years of software engineering experience, with at least a few years in a lead or principal capacity, and 2+ years hands-on with LLM or ML systems in production.
• A track record of taking an AI capability from prototype to something a client kept running - with evaluation, monitoring, and a cost model.
• Deep understanding of AI system architecture: RAG pipeline design, agentic patterns (tool use, multi-agent orchestration, memory), and when a well-tuned prompt or a plain rules engine beats either.
• Strong Python, and hands-on experience with orchestration frameworks (LangChain, LangGraph, LlamaIndex) and LLM APIs (OpenAI, Anthropic, Azure OpenAI, Bedrock) plus their trade-offs.
• Working knowledge of vector databases and semantic retrieval, and the judgement to know when retrieval is the wrong answer.
• Solid grounding in distributed systems and API design - AI is the domain, not a shortcut around the fundamentals.
• Experience with MLOps practice: versioning, experiment tracking (MLflow, Weights & Biases), evaluation frameworks, and deployment automation.
• Hands-on experience with at least one major cloud platform (preferably Azure), including identity, security, and containerised workloads at scale.
• Daily use of AI coding assistants and agentic tools in real production codebases, with a clear sense of their limits.
• Proven consultancy skills: stakeholder management, difficult conversations, and clear recommendations under pressure.
• Fluent Dutch, professional English.