Senior MLOps Engineer

Description

Our Client is the creative transformation company. We use the power of creativity to build better futures for our people, planet, clients and communities. 

In a world where media is everywhere and in everything, we bring the best platform, people, and partners together to create limitless opportunities for growth.

Global data products and technology company. We’re on a mission to transform marketing by building the fastest, most connected data platform that bridges marketing strategy to scaled activation.

We work with agencies and clients to transform the value of data by bringing together technology, data and analytics capabilities. We deliver this through the Open Media Studio, an AI-enabled media and data platform for the next era of advertising.

We’re endlessly curious. Our team of thinkers, builders, creators and problem solvers are over 1,000 strong, across 20 markets around the world.

Requirements

Essential

• GCP (esp Vertex, BigQuery, Model Registry)
• Python (needs to be very strong)
• Docker
• FastAPI (or similar)
• SQL
• Terraform
• Strong understanding or even hands-on ML experience
• Package management (uv is cool, but don’t mind others e.g. poetry; we work with a whole range here)
• Scalable experimentation & model tracking (no specific tech as we’re using native GCP logging and metadata store atm, but as we mature open-source tech e.g. MLFlow will be brilliant), and Scalable ML deployment experience (i.e. standing up an inference endpoint for hardly no traffic doesn’t count)

Desirable:

• FTI (Feature/Training/Inference) framework
• Common ML frameworks (such as PyTorch, Sklearn)

Bonus:

• RAG
• LLM orchestration tool e.g LangGraph,
• Reinforcement Learning tools e.g. OpenaiGym, RLib

Job responsibilities

The Machine Learning Engineer is responsible for deploying and maintaining the algorithms developed by data scientists. This role will be part of the Optimize Data Science team (3 FTEs plus seconded team of 5 FTEs)

While Data Scientists focus on research and model development, the ML Engineer is responsible for the technical infrastructure, scaling, performance optimization, and maintenance of the models. Their work involves implementing, testing, deploying, and monitoring the models in a production environment.

• Model deployment: Takes models (brand new or improvements to existing models) developed by data scientists and builds the software and infrastructure to deploy them into a live production environment.
• Performance optimization: Optimizes code for latency and efficiency across different hardware, like CPUs and GPUs, and overall quality (e.g. readability, maintainability, reliability and so forth).
• System design & scaling: Designs, builds, and maintains the technical components that integrate into existing software to train, deploy, and scale ML models. (N.B. There will be opportunity to work on new systems from ground up later in the year, too.)
• Monitoring and maintenance: Implements logging and monitoring to track model performance, identifies and fixes bugs (in collaboration with wider teams if appropriate), and performs necessary updates and improvements.
• Collaboration: Works closely with data scientists and wider engineering teams to understand the model and help convert it into a production-ready system.
• Experimentation: Design, build and deploy technical components based on the methodologies designed by Data Scientists to enable scalable experiments, model evaluation and visualisation of results.
• Up-skilling DS in ML Engineering and AI innovations: Support DS team to utilise modern and cloud-based (esp GCP) technologies for development (e.g. Vertex, Docker, BigQuery, Dev Containers, Ray and others) to expedite and innovate the entire development lifecycle –esp when moving from dev into prod