Review on Exceptional promise evidence for AI/ML Engineer

Stage 1
1

Hello everyone,

I am preparing my UK Global Talent Visa application under the Exceptional Promise route in Digital Technology. My profile is focused on AI/ML Engineering.

I would appreciate honest feedback on whether my evidence mapping is strong enough and whether any evidence appears weak, duplicated, or better suited for another criterion.
Proposed Evidence Structure

Mandatory Criterion – Recognition as an Emerging Leader in Digital Technology

MC01 – Institutional Recognition for a Nationally Deployed AI Product
Evidence includes an official government-backed product launch endorsed and announced jointly by a major global technology company and a national government ministry. The product serves millions of potential users across multiple languages and government domains. Evidence includes official launch announcements, press coverage by third-party media outlets, photographs taken at the official launch event, a letter from my employer’s CEO confirming my role as the sole AI engineer who designed and built the system, and screenshots of my GitHub commit history and contribution graph showing my actual code contributions to the AI components, further validating the CEO’s confirmation.

MC02 – Nationally Recognised Hackathon Award
Evidence includes a top-3 placement in a nationally recognised AI hackathon with over 5,000 applicants. Evidence includes the official results certificate and the competition page showing the applicant pool size.

MC03 – Rapid Career Progression: Promoted from Trainee to AI Team Lead in 12 Months
Evidence includes my original employment letter and official promotion letter showing progression from AI Engineer trainee to AI Team Lead within one year of joining, based on the delivery of the nationally launched AI product above.

MC04 – Open Source AI Model Contributions Recognised by the Community
Evidence includes 20+ fine-tuned AI models published publicly on an open source platform covering LLMs, TTS, ASR, and multilingual translation models for underserved language communities. Evidence includes profile screenshots showing total download counts across all models and post by engineers who have publicly referenced or used these models in their own work.

Optional Criterion 2 – Contributions Beyond Immediate Occupation

OC2-A – Personal AI Products Built and Deployed Independently Outside Employment
Evidence includes several independently built and deployed AI products I created entirely outside my employment, serving users in different sectors. Evidence includes screenshots of the live products, links to the deployed applications, and any available user metrics.

OC2-B – Open Source AI Model Contributions
Evidence includes the same open source profile referenced in MC04 but presented from a different angle: here the focus is on the act of contributing to the open source community outside my employment, the breadth of languages covered including underserved language communities, and the community impact rather than the recognition itself.

OC2-C – Community Teaching and Thought Leadership
Evidence includes a documented session teaching AI and machine learning concepts to students outside my employment. Evidence includes photographs from the session, a video recording showing me actively teaching, written feedback and testimonials from students who attended, an invitation letter from the institution, and attendance numbers.

Optional Criterion 3 – Significant Technical Contribution as an Employee

OC3-A – Technical Contribution to a Second Production AI Platform Used by Thousands of Students
Evidence includes screenshots of a second live AI platform I built during my employment, which is actively used by thousands of students. Evidence includes a letter from a senior engineer at my company confirming my specific role in building the product and the technical decisions I made, screenshots of my GitHub commit history and repository stars showing my direct code contributions to the product, and usage metrics demonstrating the scale of adoption.

OC3-B – AI Pipeline and Backend for a Third Production AI Platform
Evidence includes screenshots of a third live AI platform I built the backend and data pipelines for during my employment, accompanied by an employer letter confirming my specific technical contributions to this product.

OC3-C – Employment Contract, Promotion Letter, and Career Progression
Evidence includes my original employment contract showing my starting role, my official promotion letter to AI Team Lead, and salary progression documentation, demonstrating significant career recognition within a product-led AI company.

Recommendation Letters

I plan to submit three recommendation letters:

Letter 1 – CEO of the company I work for. This is the same person who will write the employer verification letter for MC01.

Letter 2 – A UK-based academic (a doctor and lecturer in machine learning and AI) who taught me through a certified programme.

Letter 3 – A senior academic (a doctor) who taught me machine learning and AI, with strong standing in the academic and research community.

My main concerns are:

  1. Is my overall evidence structure strong enough, and are there any obvious gaps or weaknesses I am missing?

  2. Does my recommender lineup look balanced, and is having my CEO appear in both a recommendation letter and an evidence document a problem?

Any honest feedback is very welcome. Thank you.

1 Like
2

The overall structure has enough pieces but leaks value through duplication. MC04 and OC2-B are the same open source profile presented twice, and MC03 and OC3-C are both the employment contract plus promotion letter. Assessors do not accept the “different angle” framing and will treat it as one piece counted twice, which weakens both criteria. Pick the strongest home for each and use the second slot for something new.

The self-documentation pattern runs through MC01, OC3-A and OC3-B. GitHub commit screenshots and contribution graphs are your own record of your own work, so they carry no independent weight. The employer letter is what validates your role; the screenshots read as redundant filler. If you need corroboration beyond the CEO letter, get a second senior engineer or an external technical stakeholder to write independently about the codebase.

On the CEO question: having the same person write both a recommendation letter and the MC01 verification letter is a real problem. Assessors flag templating and common authorship between reference documents routinely, and stacking one person across two document types compounds that risk. Move the MC01 verification to a different senior person at the company or at the ministry, and keep the CEO on the recommendation letter.

Both academic referees taught you through a certified programme, so they sit inside the same educational relationship. The Official Guide expects recommenders recognised in the digital technology sector who can speak directly to your contribution. Keep one academic if their standing is strong, and swap the other for someone senior in industry who has worked with you.

2 Likes
3

@Yusuf_AI

It appears you are trying very hard to make your narrative align with the criterion. You designed an AI system for a government institution; MC is not about role competence, in your context, about how you led the growth of a product‑led digital technology company, product, or team inside a digital technology company, . This has to be done in a product‑led, digital technology company, non-profit organisation or social enterprise not a government public‑sector institution. So, I am scared that it will also not rightly work for “You led the growth of a non-profit organisation or social enterprise with a specific focus on the digital technology sector” and then again, MC is about leadership, not contribution.

Recognised Hackathon Award for doing what exactly? Did you build something better than the other 5k applicants or solve a problem? Career projection or a promotion letter will not show externally validated recognition for your expertise in MC.

Open Source AI Model Contributions: This needs to be substantial, and the adoption level should demonstrate high impact.

OC2‑A is about advancing the tech sector outside your paid job, not what you built that is serving users in different sectors. It also has to be open‑sourced so an average person has access to it.

OC2‑B is the same as MC04, reusing evidence will weaken your application as evidence needs to be unique.

OC2‑C can work, but you need to clearly show you were invited to do it, so show the invitation first, what you did, and how many people were impacted.

OC3‑A can work but should be done in a product‑led company. GitHub commit history and repository stars showing your direct code contributions to the product, along with usage metrics demonstrating the scale of adoption, supported by a letter, can work, but it depends on the context of what you are actually presenting and how they align.

OC3‑B is similar to OC3‑A, you may need to present them together.

OC3‑C – Employment Contract: This criterion does not require career projection. You can use it to validate that you work for the company and show your position, which will clearly give them an idea of what you contributed based on the responsibilities tied to such roles, this are validators not evidence and you don’t want to use them to occupy space that can be used for evidence that rightly align to the criterion.

On your LOR: The CEO of the company can be okay. But for the other two academics, you need to confirm that they are experts in the sector, as it appears they are heavily academic, which means they are educational professionals and not necessarily tech experts.

On your main concerns:

  1. Is my overall evidence structure strong enough, and are there any obvious gaps or weaknesses I am missing?

You have very few listings that are okay, though too descriptive, but overall not a strong application set.

  1. Does my recommender lineup look balanced, and is having my CEO appear in both a recommendation letter and an evidence document a problem?

It could be, if the academics are experts. Having the same person can be okay if they write based on what is required and in different contexts. But I usually suggest applicants get letters from different individuals to have different voices.

All the best.

2 Likes
4

Thank you Raphael for the feedback. I really appreciate

A few clarifications:

Regarding MC01 — my employer is a private product-led company, not a government institution. The government awarded my company a contract to build the product, which is free for all public users. I led the AI engineering aspect of the product. Does it now fit the criteria of leading growth of a product-led digital technology company or product?

Regarding MC02 — the hackathon was focused on building an AI solution to a real-world problem affecting people, not a general skills competition. Does this strengthen it as MC evidence? It is open source.

Thank you.

5

Thank You @Akash_Joshi
Regarding the promotion letter appearing in both MC03 and OC3-C — I will remove it from OC3-C. Is the promotion letter strong enough on its own to anchor MC03, or does it need a supporting piece alongside it?

6

You are welcome.

Well! The fact that your company built a free for all to use product does not make it a product led company, except you have a different product that meets the product led requirements, but using this very product as a reference may still invalidate your claim. Also, you need to be clear on how you led the engineering aspect of the product.

For the hackathon, did you win, were you the one who built the AI solution, what was the entry, acceptance criterion these are some of the things that determine the strength of an evidence.

2 Likes
7

Based on what you’ve described, the profile appears to have several potentially strong elements such as strong technical ownership evidence, open-source contributions, and technical activities outside your day job for an Exceptional Promise application. However, the overall strength of an application depends not only on the quality of the evidence itself but also on how each piece is mapped to the relevant Tech Nation criteria.

MC01 is the closest fit to the first criterion, but only if you frame it as evidence that you led the growth or success of a product-led digital technology product or team, not just that you built it. If the evidence mainly shows technical development rather than growth, then it is stronger as proof of significant technical contribution than as proof of leading growth.

MC02 is useful supporting evidence, especially if the hackathon is genuinely competitive and externally recognized. If possible, include details such as the number of applicants, selection rate, judging process, and media coverage, if available, to strengthen its credibility.

MC03 is where you would be slightly cautious. Promotion alone is often considered supporting evidence rather than standalone proof of recognition. I would suggest not relying on it heavily unless you can clearly demonstrate that the promotion resulted from exceptional impact and was significantly faster than normal progression within the company.

MC04 is strong if you can demonstrate actual community adoption. Download numbers help, but independent validation is even better.

OC2-A -This section could be strengthened if you add user numbers, growth metrics, testimonials, GitHub activity, or revenue data.

OC2-B- This is where you may face a bigger risk. Using the same open-source evidence in both MC and OC2 can sometimes look repetitive. If you have enough additional evidence, I would consider replacing this with something entirely separate to reduce overlap.

OC2-C is helpful, but one teaching session on its own may be considered weak unless the audience size and impact are substantial. So, I would suggest you add more evidence such as multiple sessions, mentoring activities, workshop series or community involvement, if possible.

OC3-A appears strong as per your evidence.

OC3-B is also relevant, particularly if you can clearly explain the technical complexity and business impact of your work.

OC3-C- This is the one you need to reconsider. You already use promotion evidence under MC03. Reusing it again under OC3 may weaken the overall portfolio by creating duplication. However, here you can show user growth, revenue impact, performance improvements, or anything that demonstrates the significance of your technical contribution.

Now, Recommendation Letters are probably the area you need to improve most. The CEO letter is perfectly fine and commonly used. However, your other two recommenders are both academics who taught you. Two lecturers who taught you may make the recommendation set feel somewhat academic rather than industry-focused. Ideally, you need to aim for someone like a senior industry leader who has worked with, used, or evaluated your work and an academic or recognized expert to create a more balanced set of perspectives.

Good luck.

8

Thank you Rakesh, this is very helpful.