Lead AI Engineer, Agentic Systems - Capco AI Lab

capco; France; Sur place; EUR 120 000 - 180 000; Plein temps

Résumé du poste

Capco is seeking a Lead AI Engineer to join the AI Frontier Lab in Paris. You will own end-to-end architecture for agentic AI solutions, guide engineers, and lead client discussions on AI architecture and governance. The role combines hands-on coding with senior advisory responsibilities and client-facing delivery.

The ideal candidate has 8+ years in software/data engineering, deep cloud/platform knowledge, and proven ability to influence stakeholders and raise coding standards across the team.

Qualifications

• 8+ years in software or data engineering, including genuine enterprise architecture experience. You've designed systems that had to survive security review, integration hell and real production traffic.
• You still write code, and you'd be comfortable opening a pull request in your second week. Strong Python or equivalent.
• Demonstrable agentic AI experience: agentic loops and orchestration, tool use and function calling, multi-agent patterns, MCP or comparable integration approaches, and, critically, how you evaluate AI solutions.
• Depth on at least one major cloud platform, with the surrounding discipline: containers, IaC, observability, security.
• Presence with clients. You can hold a room of senior stakeholders, disagree respectfully, and explain a trade-off without retreating into jargon.
• A real interest in making other engineers better. This role is measured partly by what the team can do without you.

Responsabilités

• Own the architecture of agentic solutions end to end: orchestration, tool calls, memory, retrieval, observability, and budgets.
• Guide engineers through design reviews, pairing and code review; raise the bar.
• Set standards: reference architectures, guardrails, testing and eval strategy, CI/CD for non-deterministic systems.
• Make build/buy/adapt calls on frameworks, models and platforms, and defend them later.
• Prototype fast, build the hard parts yourself, push prototypes to production-grade, ensure reliability and monitoring.
• Act as senior technical voice in client discussions, translating business outcomes to system design.

Connaissances

Python

Agentic AI

Cloud platforms

Production code

Client communication

Outils

Containers

IaC

Observability

Security

Description du poste

We're looking for a Lead AI Engineer to join Capco's AI Frontier Lab, based in Paris.

The Lab builds agentic AI solutions for clients in financial services and energy, and advises them on AI architecture and governance. The engineering team is already in place. What's missing is a senior technical lead who owns the architecture, reviews and shapes the team's work, still writes code, and can represent those technical decisions directly to the client.

What you'll do

• Lead the engineering (~40%)

  Own the architecture of our agentic solutions end to end: orchestration, tool and function calling, memory and state, retrieval, evaluation, observability, plus the cost and latency budgets that decide whether a system survives contact with production.

  Guide our engineers through design reviews, pairing and code review, raising the bar rather than quietly doing the work for them.

  Set the standards: reference architectures, agentic patterns, guardrails, testing and eval strategy, CI/CD for non-deterministic systems, deployment topology.

  Make the build/buy/adapt calls on frameworks, models and platforms, and be able to defend them a year later.

• Stay hands‑on (~30%)

  Prototype fast, and build the hard parts yourself: the tricky orchestration, the eval harness, the integration nobody else wants to touch.

  Take promising prototypes to production grade, covering reliability, monitoring, versioning, rollback and human‑in‑the‑loop where it matters.

• Work with clients (~30%)

  Act as our senior technical voice in solution discussions, workshops and pitches, translating between business outcomes and system design in both directions.

  Advise on AI architecture, platform strategy, model and vendor strategy, and governance as they actually apply to a client's operating model.

  Produce content that carries weight: reference architectures, points of view, whitepapers, and talks at client forums and conferences.

What we're looking for

Essential

• 8+ years in software or data engineering, including genuine enterprise architecture experience. You've designed systems that had to survive security review, integration hell and real production traffic.
• You still write code, and you'd be comfortable opening a pull request in your second week. Strong Python or equivalent.
• Demonstrable agentic AI experience: agentic loops and orchestration, tool use and function calling, multi‑agent patterns, MCP or comparable integration approaches, and, critically, how you evaluate AI solutions.
• Depth on at least one major cloud platform, with the surrounding discipline: containers, IaC, observability, security.
• Presence with clients. You can hold a room of senior stakeholders, disagree respectfully, and explain a trade‑off without retreating into jargon.
• A real interest in making other engineers better. This role is measured partly by what the team can do without you.

Valuable additions

• Domain experience in financial services or energy, and an understanding of what regulation means for system design in those sectors.
• A consulting or client‑facing delivery background.
• Open‑source contributions, publications, or a speaking track record.
• Hands‑on AI governance, risk or assurance work.
• French alongside fluent English.

What we offer

A Lab with a real mandate and real clients. The work ships, and you'll see it running.

Influence over technical direction from day one, not after a two‑year wait.

The rare combination of deep engineering and senior advisory work, without having to give up either.