Global Talent visa for data scientists and machine learning engineers
How data scientists and machine learning engineers are assessed at Stage 1: Tech Nation's technical applicant rules, the criteria in data science terms, employer type, letters, and when a research endorsing body may fit instead.
- Prepare evidence and lettersUsually the longest partWeeks to months
- Apply online to the Home OfficeEndorsement fee£561
- Endorsing body decidesDepends on the endorsing body2 to 8+ weeks
- Apply for the visa£205 visa fee plus healthcare surchargeWithin 3 months
| Point | Tech Nation (digital technology) | Peer review (Royal Society or Royal Academy of Engineering) |
|---|---|---|
| Typical fit | Builds or leads data and ML products and platforms | Active researcher in an academic, industry or government research institution |
| Entry condition | Recognition in digital technology in the last 5 years | Active researcher with a PhD or equivalent research experience |
| Criteria | Mandatory criterion plus 2 of 4 optional criteria | Assessed by peer review against the endorsing body's guidance |
| Letters | 3 letters from established digital technology experts | 1 letter from an eminent person living in the UK, plus a second letter for Exceptional Talent |
Endorsing body
Tech Nation usually assesses data and ML roles
Tech Nation is the endorsing body for digital technology. Its published list of typical technical skills includes "Experienced data scientists/data engineers" and "Artificial Intelligence, Natural Language Processing and Machine Learning experts (AI, NLP, ML)". That puts most data scientists and machine learning engineers in the technical applicant group. The list is not complete. It gives an indication of the typical applicants Tech Nation assesses.
- Technical applicants must show proven technical expertise with the latest technologies in building, using, deploying or exploiting a technology stack and building technical infrastructure.
- Tech Nation treats you as a technical applicant if your occupation requires you to write code or manage developers.
- Exceptional Promise applicants are likely to have less than 5 years of experience in technology.
Research routes
When research-led work may fit a research endorsing body
The GOV.UK disciplines list for science, engineering, humanities and medicine places "Artificial Intelligence", "Machine Learning" and "Statistics and Operational Research" under the Royal Society. The Royal Academy of Engineering lists "Data science", "Artificial intelligence" and "Machine learning". These endorsing bodies cover active researchers in academic, industry or government research institutions. If you mainly publish research rather than build products, one of these routes may suit you better. The peer review route asks you to be an active researcher, to hold a PhD or equivalent research experience (which can include industrial research), and to work in an approved research discipline.
- Peer review needs a letter of recommendation from an eminent person living in the UK who knows you and your work.
- Exceptional Talent applicants also need a second letter from a senior member of a UK organisation, who does not need to know you personally.
- These letter rules are different from Tech Nation's three letters, so check them early.
Mandatory criterion
Recognition as a leading talent, in data science terms
The mandatory criterion asks you to show you have been recognised by others as a leading talent (Exceptional Talent) or a potential leading talent (Exceptional Promise) in digital technology in the last 5 years. Tech Nation's guide adds that all activity you give as evidence for this criterion should have occurred within the past 5 years. For a data or ML professional, the published examples that commonly fit include the ones below. They are examples only, not a checklist that guarantees endorsement.
- Leading the growth of a product or team inside a digital technology company, shown by letters from leading industry experts, news coverage or public code.
- Leading or contributing significantly to a substantial open source project outside your day job, shown by commit summaries, stars or download statistics.
- Nationally or internationally recognised prizes or awards specifically in digital technology.
- Speaking at high-profile digital technology events or specialist events in your field, with attendance estimates where possible.
- Material in professional or major trade publications about you and your work.
- A significant expert role assessing the work of others in your field.
Reading the rules is one thing. Knowing how they apply to your own evidence is another.
Book a free 30-minute callOptional criteria
The four optional criteria for data and ML work
You must meet 2 of the optional criteria, with at least 2 pieces of evidence for each. The published guidance shows five bullets, but the Immigration Rules set out four criteria. Founder innovation and employee innovation sit in one criterion, so they cannot be counted as two.
Employer type
Where you work changes which criteria are open
Tech Nation defines product-led digital technology companies as businesses that provide a proprietary digital technical service, product, platform or hardware as their primary revenue source. The guide says technical applicants from non-technical organisations are eligible, so a data scientist outside a technology company is not shut out. But two optional criteria name a product-led digital technology company, so work at other employers may only support the employee limb of the innovation criterion, recognition beyond your occupation, and academic contribution.
- The guide excludes general consultancies, outsourcing and technology consultancy groups focused on process or service delivery or solutions or systems architecture.
- The published guidance lists service delivery, process delivery, outsourcing, consultancy (technical or management), ERP consultancy and systems admin as not suitable.
- Corporate roles or experience of managing large corporate teams are also listed as not suitable.
Documents and letters
Evidence limits and the three letters
You can send up to 10 documents. At least 2 must show recognition for the mandatory criterion and at least 4 must cover 2 optional criteria, 2 for each. Each piece of evidence can be up to 3 sides of A4, and you cannot use the same piece of evidence for more than one criterion. Your CV and letters do not count towards the 10.
- Three letters of recommendation from different established experts in digital technology who have known your work for at least 12 months.
- Each letter must be written for your Global Talent application and give different examples of your achievements.
- Tech Nation says letters from an immediate colleague, manager or friend are not sufficient as evidence of recognition.
- Tech Nation does not accept templates or the use of AI tools to create letters, the personal statement or evidence.
Common mistakes
Mistakes that weaken data and ML applications
Tech Nation publishes common reasons for non-endorsement and repeats several evidence rules across its guide. These are the ones data and ML applicants most often run into.
- Evidence created only to support the timing of the application, such as a burst of GitHub activity, papers or patents just before applying.
- Showing team or company results without showing your personal work.
- Relying on internal company awards, LinkedIn or Medium posts, which Tech Nation says are not sufficient for the mandatory criterion.
- Using the same evidence for innovation and significant contribution. Tech Nation says innovation and impact are different tests.
- Mentoring only through online platforms such as ADPList, which Tech Nation says is not sufficient.
- Outdated evidence, or vague and generic letters.
Related guides
Keep reading
Guides applicants usually read next on this topic.
The UK Global Talent visa: what it is and how Stage 1 endorsement works
The Global Talent visa needs no job offer or sponsor and lets your family join you. Stage 1 is the endorsement application, where an endorsing body decides whether you are a leader or a potential leader in your field.
Read related guideGlobal Talent endorsement requirements: what each endorsing body asks for
Each endorsing body sets its own Stage 1 tests, so the requirements for a software founder, a researcher and a theatre director are very different.
Read related guideGlobal Talent endorsement documents: a checklist for each endorsing body
Each endorsing body has its own document rules, from page limits to who can write your letters.
Read related guideApplying from outside the UK. This guide covers the Stage 1 endorsement for data and machine learning roles under UK rules as published on 17 October 2026.
Article FAQ
Questions to settle before you spend time or money.
Use these answers to check the practical risks, documents, and decisions connected to this topic.
Is a data scientist a technical or business applicant?
Usually technical. GOV.UK's list of technical skills covered by Tech Nation includes experienced data scientists and data engineers, and AI, NLP and ML experts. Tech Nation treats you as technical if your occupation requires you to write code or manage developers.
Can I apply if I work at a bank or retailer rather than a tech company?
Tech Nation's guide says technical applicants from non-technical organisations are eligible. But two optional criteria name a product-led digital technology company, so fewer criteria may be open to you. Plan which two criteria you can evidence before you start.
Do Kaggle-style competition results or online courses count?
Tech Nation's guide does not list them as examples. It says your own participation in courses and formal education is not sufficient for the recognition beyond your occupation criterion. Recognised prizes specifically in digital technology are listed as possible evidence for the mandatory criterion.
I have published ML papers. Should I apply to the Royal Society?
It depends on your work. The Royal Society and the Royal Academy of Engineering cover active researchers, and their disciplines include machine learning, artificial intelligence and data science. If you mainly build products, Tech Nation is more usual, and papers may support its academic contribution criterion.
Does a patent application count for the innovation criterion?
No. Tech Nation's guide says patent applications are insufficient and all patents used as evidence must be granted, with a verifiable ID on Google Patents.
Stage 1 help
Check how your data or ML profile maps to the criteria
We help only with the Stage 1 endorsement. We can look at your work, decide which endorsing body fits, and plan evidence for the mandatory and optional criteria.
Which field is your work in?