Feedback on Evidence for exceptional promise(1st)

Stage 1
1

Hi everyone,

I’m looking to apply for the exceptional promise stage 1 soon and would genuinely appreciate feedback on the structure and distribution of my evidence.

Education History
Bsc Computer Information Systems
Msc Artificial Intelligence and Data Science

Mandatory Recommendation Letters (3):

  1. CEO of a UK AI technology company (PhD, 24 years experience, AI platform deployed across 3 continents) advisory relationship on ESG AI system
  2. Divisional Head of Data Analytics at Interswitch Group direct line manager for 16 months
  3. University Supervisor, PhD, Google Women Techmakers Fellow, L’Oreal-UNESCO Women in Science Awardee

Mandatory Criteria (2 pieces):

  1. Article published in techUK Photonics Insight Series referenced by the Council on Science and Technology in formal advice to the Prime Minister, 956 page views, 71,000 LinkedIn audience
  2. Significant contribution to SHAP (25.3k GitHub stars) merged PR acknowledged by maintainer: “had this a long time coming”.

OC2 Recognised work outside employment (2 pieces):

  1. Fairlearn PR #1614 (Microsoft’s fairness-aware ML library) expanded equal opportunity and added new False Positive Rate Parity section (112 additions), peer acknowledged by community member
  2. Combined voluntary community engagement co-facilitated Leadership, Governance and Ethics group at OpenUK AI Unconference (Oct 2025) + invited lightning talk on AI bias at AI Signals #25 (Sept 2025)

OC3 Significant technical contribution (2 pieces):

  1. CyberSource and MPGS revenue analytics dashboard at Interswitch Group Tableau/SQL, 7 stakeholders across 7 divisions, ~4M naira transaction revenue monthly, supported by PM letter
  2. Product design documentation of the above evidence

Just want to get an independent review of how my application looks overall and any suggestions to strengthen it. Thank you.

@Akash_Joshi @Raphael

1 Like
2

MC looks solid. The techUK article cited in formal advice to the Prime Minister is strong external recognition, and a merged PR on SHAP with maintainer acknowledgement is the right kind of open source contribution for Promise. Both are externally verifiable.

OC2 is reasonable. The Fairlearn PR adds to the open source track record, and the OpenUK co-facilitation plus the AI Signals lightning talk show voluntary engagement beyond employment. Make sure both the unconference and the speaking engagement are documented with organiser confirmation and audience context, not just your own description.

OC3 is where I’d tighten. You have two evidence slots used on the same project: the dashboard itself and its product design documentation. Product docs without external validation are self-documentation. Either consolidate the dashboard evidence into one piece with the PM letter as supporting material, or replace the design doc with a distinct second contribution that has its own external proof of impact.

You’re at 6 pieces across MC, OC2, and OC3. You can use up to 10, so there’s room to strengthen OC3 with a genuinely separate technical contribution if you have one.

3

@Oluebubechi_Anyahara

The authors of your letters seem generally fine. However, a university supervisor may not be suitable. Even though she is an established expert, the nature of your relationship can weaken the strength of the letter. A field based research supervisor can work, but not a university academic supervisor. This is because authors are expected to explain how they know you and the work you did with or for them. in the sector.

On your evidence sets

Your article referenced by the Council on Science and Technology can be okay, but you need to present it strategically. You should show that you were approached to write it, or that what you wrote went through a review process not like a casual tech platform article. The page views are low, and simply saying it was referenced to advise the Prime Minister may not demonstrate tech relevance. The narrative should instead focus on how your solution or insight addresses a national problem, and how the Council referenced it accordingly.

Your contribution to SHAP can work, but you must clearly state the extent of your contribution. The 25.3k GitHub stars are project level metrics, they show you contributed to something widely used, but you still need to explain why your contribution is significant. Show ownership. A PR being acknowledged demonstrates recognition, so clearly highlight that.

The Fairlearn PR #1614 can be okay, but it may be difficult to show how it significantly advanced the sector. It can support a strong OC2 evidence. The “combined voluntary community engagement” description is vague, what evidence are you presenting to show how you advanced the sector outside your paid job?

For OC3, it appears to me you are a data engineer or scientist, you need to show how you technically contributed to a product in a product‑led company. The PM’s letter can then mention the commercial outcomes resulting from your contribution. A revenue analytics dashboard or Tableau/SQL work will not demonstrate this. You need measurable, product level impact like architectures you designed, pipelines you built, performance improvements you delivered, and how these enabled product growth, reliability, or commercial success. Evidence must be traceable, specific, and externally verifiable through a platform, organisation, or credible individuals.

I think you have some good pieces, but you still need stronger evidence across the criteria to increase your chances.

All the best.

4

Your evidence sound solid however a lot depends on how these are presented in the limited number of pages. In addition to what others have mentioned, my advise for strengthening For OC2:

1. Fairlearn contribution (PR #1614)
This is strong, but you should make the impact clearer for reviewers who may not be familiar with the domain. For example: what changed for users as a result, and why it matters in practice (beyond the number of additions). If possible, also include a reference letter from a well recognised maintainer or repository owner to further validate impact and contribution.

2. Community engagement + speaking

For the OpenUK AI Unconference (co-facilitating the Leadership, Governance & Ethics group) and the AI Signals #25 lightning talk, focus on making the evidence legible and verifiable: who else was involved, what the audience size was and your position in the programme. Also keep in mind that evaluation panels tend to weight full talks and curated speaking slots more heavily than lightning talks so framing and context here really matter.

5

Thank you for your feedback, will restructure them.

6

Thank you so much for your feedback. This will really help me strengthen my evidence and present everything more clearly.