Machine Learning Engineer

Xist4 IT Limited.; City Of London; À distance; GBP 65 000 - 85 000; Plein temps

Résumé du poste

Xist4 IT Limited. seeks a Machine Learning Engineer to work on production ML across training, evaluation and inference, with responsibilities spanning data pipelines to incident debugging.

This full-time, remote role in the UK offers £65,000 to £85,000 and the chance to ship real systems that impact users.

You will collaborate with senior ML engineers and product teams in a small distributed group, with ownership growing as you ship changes and debug real-world deployments.

Qualifications

• Hands-on machine learning work beyond coursework or tutorials.
• Python used to build machine learning systems or components.
• PyTorch or JAX used for training, fine-tuning or inference.
• A model you have trained, fine-tuned or deployed yourself.
• Evaluation or testing used to understand model behaviour.
• Data or training pipelines you have built or maintained.
• Code written with production use in mind, not only notebooks.

Responsabilités

• Models: Build and improve ML components across training, evaluation and inference.
• Evaluation: Implement tests and evaluation work to understand model behaviour and catch regressions.
• Data: Build and maintain pipelines for real-world and synthetic data; ensure reproducibility.
• Production: Ship changes, observe outcomes and debug model issues, latency, cost, reliability and safety.

Connaissances

Hands-on ML

Python

PyTorch

JAX

Model training

Model evaluation

Data pipelines

Production code

Description du poste

Machine Learning Engineer | Python, PyTorch or JAX, model evaluation | Fully remote, UK
£65,000 to £85,000. Permanent.
London. Fully remote across the UK.

You will learn production ML by working on production ML. Data, training, evaluation, inference and failures are all part of the job from the start.

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 closely with senior ML engineers and product teams in a small, distributed technical group. There is guidance available, but the expectation is that your ownership grows as you prove you can ship and debug real systems.

The less glamorous work matters. Models regress, data pipelines break and production behaviour rarely looks as tidy as an experiment. You’ll be expected to investigate those problems and make the system better.

The job
Models. Build and improve ML components across training, evaluation and inference. You’ll fine-tune and adapt models as parts of larger production systems, not as isolated experiments.
Evaluation. Implement tests and evaluation work that help the team understand model behaviour. The point is to make changes measurable and catch regressions before they become user problems.
Data. Help build and maintain pipelines for real-world and synthetic data. Reproducibility matters because training loops, datasets and inference systems need to outlast the person who first wrote them.
Production. Ship changes, observe what happens and debug model issues, performance problems and incidents. You’ll work under constraints around latency, cost, reliability and safety.

What you’ll bring
Essential:

• Hands‑on machine learning work beyond coursework or tutorials.
• Python used to build machine learning systems or components.
• PyTorch or JAX used for training, fine‑tuning or inference.
• A model you have trained, fine‑tuned or deployed yourself.
• Evaluation or testing used to understand model behaviour.
• Data or training pipelines you have built or maintained.
• Code written with production use in mind, not only notebooks.
• Existing right to work in the UK.

Useful: GPU workloads, synthetic data, model serving, production incident debugging, neural architectures.

Who this suits
You are probably a Machine Learning Engineer, AI Engineer or software engineer who has moved properly into ML. You want more exposure to how models behave when real users and production constraints get involved.

It won’t suit you if you want to stop at the experiment. Here you write the code that runs in production, and you are there when it regresses or breaks.

We welcome applicants from every background and will support reasonable adjustments.