Machine Learning Engineer (Mid to Principal)

Key areas of responsibility

Key skills, experience and behaviours

Skills

• Essential
• Proven delivery: experience building, training, evaluating, optimising or deploying ML systems for real-world use, ideally in demanding environments.
• Deep domain expertise: Strong capability in at least one major area of ML, such as optimisation, computer vision, sequence modelling, LLMs, probabilistic methods, model evaluation or large-scale training.
• ML & maths depth: Strong grounding in ML/DL (optimisation, generalisation, probability, model architecture) and the ability to reason about these trade-offs in production.
• Software development: Strong Python skills and good software engineering habits, including version control, testing, code review, debugging and maintainability.
• Interpersonal skills: strong communicator who can mentor, influence, and bridge technical and non-technical audiences.
• Education: Degree, postgraduate study or equivalent practical experience in machine learning, computer science, engineering, mathematics or a related technical field.
• Builder mindset: bias to action, ownership over outcomes, and comfort working through ambiguity.
• Desirable
• MLOps excellence: reproducible pipelines, model versioning, CI/CD, observability, and automated evaluation.
• Data engineering: proficiency with Databricks, Apache Spark, Delta Lake, MLflow, and SQL; experience integrating datasets and maintaining data quality.
• Model training and optimisation: experience with pre-training, fine-tuning, distributed training, inference optimisation or adapting models for constrained environments.
• Education: PhD in AI/ML/CS or related field.
• Beneficial knowledge
• General tooling and platforms: Databricks, AWS, GCP, GitHub, Docker/Kubernetes, MLflow, Jira.
• Edge deployments: Nvidia Jetson (e.g. AGX Orin), Raspberry Pi, or other embedded accelerators.
• Distributed model training & infra: Pytorch DDP, FDSP and TorchTitan, Megatron, Slurm, Run:ai, DeepSpeed, Kubernetes, cloud or on-prem GPU clusters.
• About you
• You’ve built ML systems that persist—deployed in real settings, iterated over time, and improved through real-world feedback. You enjoy guiding others, keeping systems healthy, and making the complex understandable.

Responsibilities

• - Own and ship ML in production: take ideas from R&D to robust, maintainable deployments—often onto edge or embedded hardware. - Train and adapt models: work on model development, fine-tuning, evaluation and optimisation for real-world use cases. - Work at scale where needed: run and improve training and inference workloads across GPUs, including multi-GPU or multi-node environments, to support models that can perform reliably in constrained settings. - Improve performance: profile, optimise and debug ML systems across model code, data pipelines, inference stacks and hardware constraints. - Own evaluation quality: design evaluation pipelines, benchmarks, test sets and feedback loops that help us understand model behaviour before and after deployment. - End-to-end ownership: data collection/curation, feature engineering, model training, evaluation, deployment, monitoring, and iteration. - MLOps/LLMOps: CI/CD for models, containerisation/orchestration, experiment tracking and registry, model evaluation pipelines, safety guardrails, canaries, and performance monitoring. - Cross-team collaboration: partner with software, systems, and product colleagues; simplify complex topics for other disciplines and customers. - Data foundations: establish pragmatic data pipelines (batch/stream) that make curation, provenance, and reproducibility first-class. - Raise the bar: depending on level, mentor others, guide technical decisions and improve engineering standards across the team.