Graduate AI Engineer

Graduate Machine Learning Engineer role at Permutable developing production AI systems, large language models, and NLP solutions to deliver real-time market intelligence for institutional financial clients.

Build production AI systems that interpret global markets

Permutable is building the intelligence infrastructure institutions use to understand what is moving global markets - and why.

Our technology transforms large volumes of multilingual news, economic developments, market narratives and geopolitical information into structured, explainable signals. These signals support institutional investors, banks, asset managers, energy desks and commodity trading teams across macroeconomic, commodities, currency and geopolitical research.

Permutable currently analyses more than 250,000 sources across over 80 languages, processes more than one million narratives each day and covers more than 70 assets. In 2026, we were named Hedgeweek’s Technology Provider of the Year: Innovation.

We are now looking for an exceptional graduate engineer to help us build the next generation of our machine learning and market intelligence systems.

The opportunity

This is not a rotational graduate scheme or a role where you will spend your first year observing from the sidelines.

You will join our engineering and data science team, work on real technical problems and contribute to systems used in live institutional workflows. You will be supported by experienced colleagues, but you will also be trusted with meaningful responsibility from an early stage.

The work sits at the intersection of:

• Machine learning and natural language processing

• Large language models and agentic systems

• Multilingual information retrieval

• Time-series and point-in-time data

• Financial markets and economic research

• Scalable production engineering

You might be improving how an emerging economic narrative is detected across multiple languages one week and evaluating whether a new model can identify changes in commodity-market pressure the next. Your work will not remain in a notebook. You will help take ideas from research and experimentation through to reliable production systems.

What you will work on

Depending on your strengths and interests, you will:

• Design, train and evaluate machine learning and NLP models operating on large-scale textual datasets.

• Develop systems for classification, information extraction, entity resolution, narrative clustering, sentiment analysis and semantic search.

• Experiment with transformers, embeddings, retrieval systems and large language models.

• Build robust evaluation frameworks to measure accuracy, consistency, latency and production performance.

• Carry out detailed error analysis and translate findings into practical model improvements.

• Work with engineers, data scientists, market analysts and product colleagues to solve commercially relevant problems.

• Communicate technical findings clearly to both technical and non-technical team members.

• Contribute ideas to our research direction, product architecture and engineering standards.

What success could look like

During your first few months, you could:

• Become familiar with Permutable’s data, models and production architecture.

• Ship a contained improvement to an existing model, evaluation process or data pipeline.

• Reproduce and assess an existing experiment using historical point-in-time data.

• Present your findings and recommendations to the wider technical team.

As your experience grows, you could:

• Take ownership of a model, service or technical workstream.

• Design and run original experiments.

• Deliver new capabilities into production.

• Help determine how our machine learning systems evolve.

• Work directly with market analysts and institutional use cases.

• Mentor future graduate engineers joining the team.

What we are looking for

You may be completing your degree or have graduated within the past two years.

We are particularly interested in UK graduates from:

• Computer science

• Artificial intelligence or machine learning

• Mathematics or statistics

• Physics

• Electrical, mechanical or software engineering

• Data science

• Computational linguistics

• Another highly quantitative discipline

You should have:

• Strong Python programming skills.

• Solid foundations in algorithms, data structures and software engineering.

• A good understanding of probability, statistics, linear algebra and machine learning.

• Experience using at least one machine learning framework.

• The ability to test ideas methodically rather than relying on intuition alone.

• Evidence that you can take a difficult technical problem, break it down and make progress independently.

• Clear written and verbal communication skills.

• Intellectual curiosity and a genuine interest in understanding how systems work.

• The confidence to ask questions, challenge assumptions and defend your reasoning.

• The motivation to work in a fast-moving startup where priorities can evolve quickly.

We do not expect you to arrive knowing everything. We do expect you to learn quickly, care about technical quality and take ownership of your work.

Evidence that would make you stand out

Strong candidates may have completed one or more of the following:

• A technically ambitious dissertation or research project.

• An internship involving software engineering, machine learning or quantitative research.

• A substantial personal or university project that progressed beyond a standard tutorial.

• Open-source contributions or a well-documented GitHub repository.

• Research involving NLP, transformers, LLMs, time-series data or information retrieval.

• Competitive programming, mathematical competitions, hackathons or data-science competitions.

• An academic publication, preprint or research assistantship.

• Work involving large, noisy or multilingual datasets.

• A project where you had to make and justify difficult modelling or engineering trade-offs.

We are more interested in the depth of your thinking and the decisions you made than in an unnecessarily polished portfolio.

Useful, but not essential

Experience with any of the following would be helpful:

• PyTorch, TensorFlow, JAX or scikit-learn

• Hugging Face and transformer-based models

• Embeddings, vector search or retrieval-augmented generation

• SQL and large-scale data processing

• AWS or another cloud platform

• Docker, APIs and production deployment

• Experiment tracking and model monitoring

• Financial markets, economics, energy or commodities

• Time-series modelling

• Multilingual NLP

• Distributed systems

Previous experience in finance is not required. We are looking for excellent technical potential and an interest in learning how global markets work.