Contract - Machine Learning Operations Engineer

On-site in London·Added today (last checked 22 Oct)

Deloitte UK

630 open roles

Apply

Overview

Contract: Contract role, inside IR35

Skills

Nice to have

The role

Location

Birmingham, Bristol, Cardiff, Edinburgh, Glasgow, Leeds, London, Manchester

Business Line

Technology & Transformation

Job Title: Machine Learning Operations Engineer (contract)

Location: Hybrid - open to any Deloitte office location (London, Bristol, Birmingham, Manchester, Leeds, Cardiff, Edinburgh, Glasgow). Travel to office 3 days per week

Duration: 2 months

Contract Start Date: ASAP November 2026

Day Rate: Competitive (Inside IR35)

Deloitte

Working with the Deloitte Associate (Contractor) Programme means we can offer you the opportunity to work on a variation of industry and client related projects. Our aim is to retain the best talent and so when your project end date nears our team of Talent Community Advisors will be working with you to look at alternative projects within the firm that suit your experience should you wish to continue with Deloitte.

The Role

We are looking for an experienced Machine Learning Operations Engineer to help design, build and deploy enterprise AI and Copilot solutions for one of our Financial Services clients. You will work across machine learning, retrieval-augmented generation (RAG), agent orchestration and cloud platforms to deliver scalable, production-grade AI capabilities. The role requires hands-on experience building ML pipelines, deploying models into production, integrating enterprise data sources and implementing robust governance, monitoring and evaluation frameworks. Preferably someone from a Financial Services background.

Responsibilities but not limited to;

Design and ship production Copilot agents, owning them from prototype through production
Build retrieval that holds up in production: Graph RAG combining knowledge-graph traversal over accounts, contacts, products, and prior engagements with vector search,
Design agent memory: episodic and semantic memory organized as memory nodes, with recall, summarization, and decay policies tuned per use case,
Own organizational context management
Build and productionize ML model training pipelines
Build batch and online inference for those models
Own answer quality as an engineering discipline
Compose multi-agent workflows using A2A, and integrate tools and data through MCP servers (CRM, client and product reference data, document stores, calendar and email, Databricks, Snowflake, Kubernetes, Confluence, Bitbucket, Splunk).
Build and run evals: task-level and end-to-end agent evaluations, regression suites, LLM-as-judge, human review loops with sellers, and quality and safety gating before release.
Close the loop with production feedback
Produce the model documentation, test evidence and controls
Deploy and operate solutions on public cloud (AWS and/or Azure) with strong SDLC, security, resiliency, and observability practices.
Partner with sales, coverage, product and business partners to turn use cases into shipped, supported agents.

Essential Skills and Experience

> 4 years experience building Copilot or AI agent solutions preferably from a Financial Services background
> 4 years experience building and deploying machine learning models in production environments.
Experience of Graph RAG and MCP integrations.
Expertise in ML Ops practices, CI/CD, model monitoring and lifecycle management.
Experience with Generative AI, LLMs and Retrieval-Augmented Generation (RAG) architectures.
Strong Python development skills and experience with Databricks.
Experience developing batch and real-time inference pipelines.
Knowledge of vector search, retrieval optimisation and AI evaluation frameworks.
Experience working with AWS and/or Azure cloud platforms.
Familiarity with Kubernetes, Snowflake and enterprise data integrations.
Strong understanding of security, observability and AI governance principles.
Excellent stakeholder management and communication skills

Deliverables: Responsibilities but not limited to:

Design, build and own production-grade copilot agents that provide secure, context-aware and persistent support across the client relationship.
Deliver production ML and retrieval capabilities that turn enterprise data into timely, explainable recommendations for sellers.
Establish the evaluation, governance and operational controls required for trusted AI at enterprise scale.

IR35

As a means of managing tax, commercial and reputational risks, Deloitte prohibits the use of Associates through Personal Service Companies ('PSCs'). All Associates must contract under PAYE arrangements through a Deloitte approved 'Employment Company' (aka 'umbrella company.')