Senior Machine Learning Engineer

Xist4 IT Limited.; City Of London; Remote; GBP 95,000 - 120,000; Full time

Job summary

Xist4 IT Limited is seeking a Senior Machine Learning Engineer to own the core ML systems in a fully remote, UK-wide setup. You will manage data pipelines, training, inference, and monitoring, while collaborating closely with research, product, and engineering teams.

You will mentor peers and drive production quality in a fast-moving AI product environment. The role emphasizes building robust, scalable ML solutions, balancing accuracy with latency and cost, and owning the end-to-end lifecycle

Qualifications

• Experience shipping ML systems into production.
• Strong Python used for ML systems in production.
• Hands-on with PyTorch or JAX beyond experiments.
• Ownership of data, training, evaluation and inference.

Responsibilities

• Build core ML systems from data preparation to inference.
• Turn research ideas into production-ready systems serving real users.
• Write production-grade Python for ML pipelines and training/evaluation.
• Mentor and review work to raise engineering standards.

Skills

ML systems

Python for ML

PyTorch/JAX

Production ML

Production debugging

Mentoring

UK work rights

Tools

PyTorch

JAX

Job description

Senior Machine Learning Engineer | Python, PyTorch, JAX | Fully remote, UK
£95,000 to £120,000. Permanent.
London. Fully remote across the UK.

This is for an ML engineer who has already learned the uncomfortable bit: a model that works in a notebook is only the start. You will own core ML systems from data and training through to inference and how they behave in production.

Our client is an early-stage AI product company building applications that get on with everyday tasks before you ask. A prototype exists, launch is ahead, and the company is funded without relying on an upcoming round.

You'll work in a small, distributed technical team. The role is hands-on and the work is yours to own, with close links to research, product and engineering. You'll also review and mentor other ML engineers, mostly through the quality of your own technical decisions.

The trade-offs are real. Accuracy matters, but so do latency, cost, reliability and safety. Production failures are part of the work, and you'll be expected to trace them through models, data and systems instead of treating them as somebody else's problem.

The job

ML systems. Build the core machine learning systems for a long-horizon AI product. You'll take work from data preparation through training, evaluation, inference and iteration.

Production. Turn research ideas into systems that can serve real users. You'll investigate model failures and system issues from production signals, then ship changes and measure whether they worked.

Engineering. Write production-quality Python and work with PyTorch or JAX on GPU-based training and inference. Training, inference and data pipelines need to keep working, and stay maintainable, as the product changes.

Technical judgement. Own ambiguous problems without waiting for a detailed specification. Review work from other ML engineers and help raise the engineering standard through practical decisions, not process for its own sake.

What you'll bring

Essential:

• ML systems you have built and shipped to real users.
• Production-quality Python used for machine learning systems.
• PyTorch or JAX used beyond isolated experiments.
• Ownership across data, training, evaluation and inference.
• Production model failures you have investigated from real signals.
• Work balancing model quality against latency, reliability or cost.
• Other ML engineers whose work you have reviewed or who you have mentored.
• Existing right to work in the UK.

Useful: GPU training systems, inference optimisation, synthetic data, long-horizon models.

Who this suits

You are probably a Senior Machine Learning Engineer, Senior AI Engineer or experienced ML Engineer who prefers building and running systems to presenting research. You can take a vague problem and get it safely into production.

It won't suit you if you'd rather hand a model over once the evaluation looks good. Here you stay with it through inference, real users and whatever fails afterwards.

Let's talk

Actively looking or just curious, We welcome applicants from every background and will support reasonable adjustments.