Request for Evidence Review – Data Science / AI Exceptional Promise

Stage 1: Tech Nation Endorsement
1

Request for Evidence Review Data Science / AI (Exceptional Promise)

Hi everyone,

I’m preparing to reapply for the UK Global Talent endorsement in Digital Technology from Nigeria and would really appreciate an honest review of my evidence portfolio.

I previously applied and was unsuccessful, so I’ve spent time strengthening my evidence,

I would appreciate feedback on whether the evidence below is strong enough, whether the evidence is appropriately positioned, and whether there are any obvious weaknesses or gaps.

Mandatory Criteria

1. Leadership / significant role

  • Senior Data Scientist / Senior Data Analyst at a fintech startup.

  • Played a significant role in the growth and development of the data/analytics function.

  • The fintech startup has raised over $3.5 million in funding.

  • Evidence of my impact on the company’s products, analytics and decision-making.

  • Recommendation/reference letter from the VP of Technology and Product describing my role, responsibilities and impact.

  • Evidence of the scale and business impact of the products/analytics work I contributed to.

2. Judging / assessing the work of others

  • Judge/reviewer for the Tech4Dev Women Techsters Fellowship Hackathon

  • Reviewed research papers in the digital technology field for:

    • IndabaX Nigeria – Deep Learning

    • International Conference on Innovations in Electronics, Communication and Computing Techniques (IEC2T-26)

    • Medical Image Computing and Computer-Assisted Interventions (MICCAI-MSB EMERGE 2026). Kindly note that each of the activities listed above is supported by corresponding evidence, including letters of invitation, proof of participation/review, review confirmations, and official certificates or letters confirming my involvement.

3. Media / professional publications

  • Publication in BusinessDay newspaper relating to my work and expertise in AI/data technology.

  • Publication in The Sun newspaper.

  • Coverage/features (interviews and expert opinions) through digital media outlets including Channels TV and TVC.

  • Articles and media coverage focused on my work and expertise in data science, AI and digital technology.

4. Speaking engagements

  • Speaker at DataFestAfrica.

  • Speaker at PyCon Nigeria.

  • Speaker at an AWS West Africa conference/event.

  • Speaker at a CCHub machine learning event.

  • Other technology-focused speaking and educational engagements.

OC 2

1)Open-source contributions

  • Contributions to open-source projects including:

    • Fairlearn

    • River ML

    • Evidently

  • Approximately 5 contributions on average across the open-source projects.

  • Evidence showing my contribution activity and ranking/position among project contributors.

2)Thought leadership / technical publications

  • Articles published on HackerNoon.

  • Articles published on DZone.

  • Approximately 5 articles on each platform.

  • Topics covering AI, Retrieval-Augmented Generation (RAG), machine learning and data/ML engineering. (The articles published on these platforms undergo a review and editorial process before they are approved and published.)

  • Articles have collectively received significant readership, with some averaging over 1,000 views.

3)Mentorship

  • Structured mentorship through:

    • Niyo Group

    • Google DeepMind / TRI AI

    • Break Through Tech

  • Mentored fellows/students as a data science and AI professional.

  • Evidence includes mentorship roles, letter of invitation, programmes, cohort information and feedback/confirmation of participation .

OC4

Research / Academic Contributions

Conference presentations

  • Presentation at the Frontiers in Computing for Sustainable Development academic conference at the University of Ibadan.

  • Presentation at the 10th International Conference on Cloud and Big Data Computing (ICCBDC 2026) at the University of Cambridge.

  • Presentation at the 10th International Conference on Advances in Artificial Intelligence (ICAAI), London

    • Chunking and Retrieval Strategies for Reducing Hallucination in Domain-Specific Retrieval-Augmented Generation Pipelines

Research papers

  1. Designing a Full-Stack Pipeline: From Data Collection to Edge Deployment in Low-Bandwidth Environments

  2. Crop Disease Detection Using Convolutional Neural Networks in Low-Resource Agricultural Settings

  3. Chunking and Retrieval Strategies for Reducing Hallucination in Domain-Specific Retrieval-Augmented Generation Pipelines. Kindly note that I plan to reapply in early Q2 next year, as some of my research papers are still awaiting publication, they are expected to be published in September and October.

Recommendation letter

  1. CEO of a YC tech startup (Fintech)

  2. CTO of a Non-profit tech organisation (in the UK)

  3. CTO of a fintech in Nigeria, established in the digital technology field

I would particularly appreciate feedback on:

  • Which pieces of evidence appear strongest?

  • Are any of these pieces too weak or too internally focused?

  • Does the portfolio demonstrate recognition beyond my immediate employer?

  • Are the open-source, publications, judging, speaking and research evidence sufficiently independent?

  • Are there any areas where I should obtain stronger third-party evidence before reapplying?

  • Does the overall portfolio tell a convincing story of progression and recognition in AI/Data Science?

Thanks in advance for any honest feedback.

2

@Akash_Joshi @Raphael @Francisca_Chiedu @May

3

You’ve listed far more than the cap allows. The evidence limit is ten documents total, two for the Mandatory Criteria and at least four split across your two optional ones, and your list runs well past that. The RAG chunking paper is also doing double duty under both conference presentations and research papers, which isn’t allowed since one piece can’t cover two criteria. Decide the two strongest pieces per criterion first, then cut the rest.

Recognition beyond your employer sits mostly in your externally-facing evidence. The Guide lists assessing the work of others and media about you personally as valid Mandatory evidence, so the judging and the AI coverage are your natural anchors there. The fintech leadership piece is softer, because it leans on the VP letter and the Guide discounts recognition coming from an immediate manager.

Open source is your weakest optional piece. A handful of contributions spread across three projects won’t clear the consistent and ongoing track record the Guide asks for under OC2. Either deepen one project into something substantive or drop it and carry OC2 on the thought leadership and mentorship, checking the mentorship is a structured programme with selection criteria that runs in person, since online-only programmes are explicitly excluded.

Waiting for the papers to publish is the right call. Anything unpublished, or anything that looks produced to time the application, works against you rather than for you, so reapply once they are peer-reviewed and out.

4

Based on the information you shared, the mandatory criterion appears to have useful material, particularly the media coverage, speaking engagements and judging/reviewing activities. However, the Senior Data Scientist role and the company’s $3.5m funding should mainly provide context; they do not by themselves establish recognition as potential talent. The strongest mandatory evidence would be the independent recognition through reputable media, conference speaking and judging/reviewing, provided the evidence clearly attributes the recognition to the applicant personally.
For the optional criteria, OC2 is potentially well supported, particularly through the open-source contributions, technical publications and mentoring. However, the evidence should demonstrate that these activities contributed to the advancement of the field and, where possible, include independent recognition or validation rather than relying mainly on contribution counts, publication numbers or participation. The publications would be stronger where there is demonstrable readership, technical relevance and evidence that the applicant is the author and subject-matter contributor.
The research evidence could potentially support the academic/research criterion, particularly if the papers are formally published or independently endorsed by an expert before submission. Conference presentations alone would not necessarily satisfy this criterion, so the eventual publication or expert-endorsement status of the papers will matter. The applicant should also ensure that the same evidence is not being reused across multiple criteria, as the current guidance requires separate evidence for each criterion.
The three recommenders appear directionally relevant if they are genuinely established digital-technology experts and have known the applicant’s work for at least 12 months. Their titles and organisations should establish that expertise clearly, and their letters should specifically explain why the applicant demonstrates potential leadership, rather than simply confirming employment or technical ability.

5

@Balogun_David_Taiwo Trust you’re good.

It’s good that you want to give it another shot, and it’s clear you’ve improved on your previous application. That’s already a strategic step forward.

On your MC Leadership / Significant Role, your achievements as a Senior Data Analyst are noted. However, MC is really about leading the development of a product, a team, or the direction of a technology company and being recognised for doing so, with clear evidence that those things have or can advance the sector.

Judging or assessing the work of others can work for MC, but the key question is, are these women your peers, and are they in your area of specialisation? If yes, then it may hold. Peer review of research can also meet MC, but Tech Nation usually checks whether the applicant is qualified to assess academic research or journals. Top tier papers are often reviewed by academics with research experience, and I’ve seen cases where feedback came back saying the applicant wasn’t sufficiently qualified to assess the work of others in academics. So while you can use it, be mindful of that.

On Media / Professional Publications, your media engagements can be okay, but you need to clearly show the recognition element like invitation letters, emails stating you were invited as a domain expert, and that the topic was relevant to the tech sector.

This criterion list is essentially asking you to prove that you were the subject of published coverage because of your work in digital technology, and that the publication is credible with clear details about you. So it has to be tied to something you’ve done that has impacted the sector. Just being published in a newspaper won’t hold. They usually dismiss those as advertorial or promotional Also, expert opinion fits OC2 more, where thought leadership is required. Your speaking engagements can be useful, but you didn’t list the evidence you’re submitting. Were you invited because of your expertise? Were you a keynote speaker? Can you show evidence of you speaking on stage to at least 100 people? If the videos are online, like on YouTube, can you show post event impact? Also note that they check the credibility of the conferences.

For OC2, your contributions to Fairlearn, River ML and Evidently can be okay, but they check the relevance of the project. For instace, if someone contributes a basic code snippet that predicts weather, they often get feedback that the project is too basic or not relevant for the calibre of applicants endorsed through this route. So make sure the projects are relevant, and clearly show significant, sustained contributions over time. Even five contributions can be okay if the impact and level of contribution are strong.

Thought leadership or technical publications can also work, but you need to show acceptance evidence and editorial process evidence. Beyond stating it, show proof. And 1000 views across all your articles is quite low.

Your mentorship can be useful with the evidence you’ve presented, but again, you need to show impact and sector advancement. That’s key for OC2.

On OC4, once the conferences are top tier and the papers are in top tier journals, in a relevant tech‑related area, they should be okay. But you still need to show how these have advanced the sector through citations, readership, downloads, references, viewership. All these help demonstrate impact and sector advancement.

You have some good listings, but it will really help if you’re clearer about the evidence you’re presenting.

All the best.

6

Thank you, Joshi!

Regarding the RAG chunking paper, I will use it solely as evidence of my research work rather than counting it under both research papers and conference presentations.

For the open-source work, I will definitely deepen my projects and, most importantly, focus on making them more impactful, with clear evidence of adoption, usage, or tangible outcomes.

The mentorship programmes are well-structured, but they are conducted entirely online due to geographical differences. For example, Break Through Tech involves participants across the USA and Canada, while TRI AI involves mentees across West Africa and North America. Does the fact that these programmes are fully online mean they don’t stand a strong chance as evidence, or can they still be valuable if I can demonstrate their structure, scale, impact, and outcomes?

For the TRI AI mentorship programme specifically, I mentored around 10 different groups, with an average of 5–6 mentees per group.

Also, regarding a supporting letter from my VP of Technology and Product validating my work at the startup, does the fact that he is my immediate manager significantly weaken its value as evidence? Would it be better to focus more on the work and impact itself for example, my role in building the data architecture, establishing data metrics and analytics infrastructure, and, most importantly, demonstrating how my work contributed to the startup’s growth and fundraising?

7

Thank you, @Noddy

I will make sure to clearly demonstrate the impact of my work at the startup, particularly how I contributed to building the data architecture and analytics infrastructure, which ultimately supported the development of several dashboards used to monitor key areas of company growth, including sales, revenue, and other business metrics.

Regarding OC2, the technical articles are highly relevant. Recently, an Engineering Lead/Partner at IBM shared one of my technical articles on LinkedIn, and my articles have collectively attracted significant readership, with over 1,000 readers per article in many cases and some receiving more than 3,000 readers. I will make sure to provide evidence of this reach and engagement.

For the conferences, the papers and presentations are subject to peer review, with some of the reviews and academic involvement coming from professors and researchers in AI across major university institutions. I also intend to clearly distinguish between my research publications and conference presentations and will not use the same evidence twice, as I understand the implications of double-counting evidence.

Once again, thank you so much for your guidance and advice. I really appreciate it.

8

Thank you so much, @Raphael I trust you are also doing well.

Regarding my impact at the startup, I have several pieces of evidence that demonstrate my contributions, including building the data architecture, developing analytical workflows, and creating dashboards for the organisation. These ultimately enabled the company to better track key growth metrics and supported data-driven decision-making, which contributed to the company’s growth and fundraising efforts.

Regarding Accessing the Work of Others, particularly the Tech4Dev engagement, I was invited as an expert to assess and evaluate participants’ work, and this was directly within my area of specialisation in Data and AI.

For peer review of research, I have written and continue to work on several research papers, so I believe I have developed the relevant knowledge and expertise to review research in this field. This is also why I have listed multiple peer-review activities rather than relying on a single conference review. Reviewing papers for different conferences, across different research topics within AI and Data, I believe demonstrates greater depth and breadth in my reviewing experience. I really appreciate you drawing my attention to the feedback you have seen from previous applicants, as it has helped me think more carefully about how I present this evidence.

Regarding the TV invitations, most of the initial communication for one of the engagements was informal, through WhatsApp, where the TV anchor invited me to participate. The other TV engagement was initiated through LinkedIn. Would the WhatsApp/LinkedIn chat communication still be acceptable as evidence of an invitation if I can provide the relevant screenshots and supporting evidence?

The topics I discussed on both TVC and Channels TV were directly related to my professional expertise in Technology, particularly AI and Data. The discussions covered areas connected to my research, experience working with startups, the future of AI in Africa and my teaching and mentoring work in the Data and AI space.

Similarly, my other media features in BusinessDay and The Sun were also related to my expertise in Data and AI. One article was based on my work around synthetic data, while another focused on data, climate change, and sustainability, which was particularly connected to my work at a clean-energy startup.

For the speaking engagements, I plan to use these as part of my Mandatory Criterion evidence. I have supporting evidence including invitation emails, event flyers, photographs, and evidence of my participation on the main stages, with audiences of 100+ attendees. Some of these include events such as PyCon Nigeria, AWS West Africa, and DataFest Africa, which I believe are credible platforms.

Regarding open source, I am still actively contributing to these projects. I also intend to focus on making my contributions more substantial and impactful, rather than simply accumulating a large number of minor contributions. I want to be able to demonstrate tangible outcomes, adoption, or value from the projects.

For thought leadership/technical publications, each of my articles receives over 1,000 readers individually, rather than 1,000 readers collectively. Some articles have received more than 3,000 readers. More recently, an Engineering Lead/Partner at IBM shared one of my DZone technical articles on LinkedIn. I will make sure to provide evidence of the readership, engagement, and external recognition where possible.

I will also continue to demonstrate the impact of my mentorship activities, including progression and career advancement where I can obtain appropriate evidence.

Regarding OC4, I will also work on strengthening the evidence around readership, references, citations, and downloads. Over the weekend, two additional research papers was accepted: one for the International Conference on Intelligent Technology, System and Service for Internet of Everything (ITSS-IoE 2026) at Babcock University, and another for the 3rd International Conference on Artificial Intelligence, Robotics, and Cybersecurity for Emerging Technologies in Essaouira, Morocco. I intend to include the corresponding paper presentations as part of my overall research evidence.

Finally, I wanted to ask about mentoring university students at the University of Bristol through the Bristol Data Science and Artificial Intelligence Society (BDSS). Would this be appropriate to include as part of my mentoring evidence, given that the programme involves mentoring university students? I am asking specifically because it is university-level mentoring, and I can provide evidence of the structure, my involvement, and the impact on the students.

Thank you once again for your detailed feedback and guidance. I really appreciate you taking the time to review these points, and your previous feedback has been very helpful in shaping how I approach and present my evidence.

9

What the TV appearance produced matters more than the channel the invite came through. A WhatsApp or LinkedIn screenshot on its own is thin proof, so anchor this evidence on the actual broadcast segment showing you discussing AI or data as an expert. Use the screenshots to corroborate that you were approached for your expertise. The recognition the Guide looks for lives in the aired content about you.

University mentoring can work well for OC2. The Guide lists teaching at a university and mentoring through structured programmes with selection criteria as valid examples, so Bristol BDSS qualifies if you can show the programme structure alongside your personal contribution and clear outcomes for the students. The one thing to be ready for is the in-person expectation, so make clear how much of the mentoring is delivered face to face.