Senior AI Platform Engineer - LLM Infra & FinServ

Capgemini; Greater London; On-site; GBP 90,000 - 150,000; Full time

Benefits offered by this job

Applaud recognition portal

Great Place to Work wellbeing program

Mental Health Champions

Learning for life: courses and certs

Job summary

Capgemini's NewCo, an AI-native product organisation within Capgemini Financial Services, builds platform-focused software for insurance claims and payments with a small, senior team. Engineers author specifications, tooling, evaluation suites, and guardrails, while AI agents handle much of the implementation.

As a platform engineer you will own end-to-end components, from model gateway and orchestration to evaluation harnesses and guardrails, and you will work with product teams to embed new

Qualifications

• Strong production engineering: Python and a second systems language; Kubernetes, IaC and cloud.
• Hands-on experience with LLM infrastructure: model gateway, model serving, vector/retrieval systems.
• Experience with agent orchestration frameworks and tool calling concepts.
• Familiarity with eval harnesses and CI gating for model quality.

Responsibilities

• One or more platform components end to end: model gateway, agent runtime/orchestration, evaluation harness, or guardrail engine.
• Develop training/adaptation loops: pipelines turning production traces into RLHF/fine-tuning data.
• Ensure reliability, latency, and cost efficiency of platform services with solid docs.
• Create self-service surfaces for product engineers with sane defaults and documentation.

Skills

Python

Go

Rust

Kubernetes

IaC

Cloud

LLM infra

Agent frameworks

Tools

LangGraph

MCP

Job description

Capgemini's NewCo, an AI-native product organisation within Capgemini Financial Services, builds platform-focused software for insurance claims and payments with a small, senior team. Engineers author specifications, tooling, evaluation suites, and guardrails, while AI agents handle much of the implementation.

As a platform engineer you will own end-to-end components, from model gateway and orchestration to evaluation harnesses and guardrails, and you will work with product teams to embed new