Evidence review help - Exceptional Promise

Stage 1
1

Hi everyone,

I’m preparing my Exceptional Promise (Digital Technology) application and would really appreciate feedback on whether my evidence is sufficient.

Mandatory Criteria - 1: Press Coverage
My work was addressed in press coverage - Le Monde and Radio Canada. The catch is that the PI talked about my work. However, he was kind enough to give me a separate letter of attestation stating that it was my work, which he spoke about.

Mandatory Criteria - 2: Judging of work
I have been reviewing for conferences in Computer Vision, Pattern Recognition and Graphics. I’m thinking of using NeurIPS (2025) and ToG (2024)

Mandatory Criteria - 3: Letter of Affirmation
I’m using the PI’s letter for MC-1 as the third document for MC-3.

Optional Criteria - 4A (Research): Talks
I have given talks at conferences - SIGGRAPH Asia (ToG) 2024 and ECCV 2022. They were ~10min with some 5-7min of questions. The conference was attended by 6k~ people. But I don’t have the exact figure of those in the crowd. I am using a pic of myself on the stage, slide 0 where it shows my name. It sort of looks crowded, but I’m afraid I don’t have videos or further evidence of my talk. Is that alright? I am also providing a conference public page that lists my name as a speaker.

Optional Criteria - 4B (Research): Papers

I’ve 2 papers at CVPR (first author) and one at NeurIPS (first author). I’m using them by showing the first page of the paper alongside attaching a link to the full version.

Optional Criteria - 3A (Significant Contrib): Impact of work

I’ve worked on creating a dataset and a publicly usable tracking software. These two have ~100 users. The dataset is hosted on a public benchmarking site. User count seems pretty weak, is my concern

Optional Criteria - 3B (Investments in my product):
I’ve built a novel tracking software that was both published and used in real-world applications. This tracker was a backbone in the subsequent EU projects that my PI undertook. The usage of the tracker, alongside monetary value, is mentioned in the respective project page. Total investment in the two projects equals 6.9mn Euros.

The investment, of course, doesn’t name me directly. Hence, I’ve the attestation confirming this in the letter that’s in Mandatory 3/

Optional Criteria - 3C (OSS Contribution):

Perhaps this section is my weakest! I’m using a merged PR in PyTorch, issues reported in PyTorch3D across the last 3y and my contribution to py-motmetric, which I developed during the time I built the tracker. For Pytorch, it was a trivial change and nothing very impactful (docstring edit - that’s what makes me a bit concerned). Though I only needed 2, I still included them in case the other two were weak.

Your thoughts and comments are welcome! Thanks a lot :slight_smile:

Educational background: B.E, MsC, PhD.

EDIT:

LoR-1: A Professor from UCL
LoR-2: A Professor from Ecole Polytechnique (Paris)
LoR-3: A Research Director at INRIA (France)

Best,

2

Hi @srs04 you have a solid set of evidence outline! Some refinement can lead you to a strong shot at this.

MC1: what evidences are you showing here for the two? Is it going to be screenshots of the interview? Can you clarify the catch again - it looks like the PI covered your work with your name and also shared a letter?
Since MC3 is connected to MC1, it’s better to combine both into one evidence document.
MC2: this is a great evidence! Both conferences look credible, if you have more feel free to add them. Ensure you establish the stature and scale of the conferences in tech.
OC4A: conferences are great! You don’t need to mention the details of the time of your speech. Is it possible for you to show an invite letter inviting you as a speaker to these conferences? Possible to get a reference letter confirming the audience? Any PR or news coverage about this conference that affirms the scale? OC4B is a strong evidence too.
OC3 looks like a weak section as it needs you to demonstrate quantified impact of your work. <100 users is too low. You can’t reference a letter of another criteria, you will need another letter for the second one. It’s a good idea to have a letter here to link the PR to your contribution alongwith the quantified impact. The third one is also not strong enough to clear.

Do you want to evaluate if you can attempt OC1 or OC2 instead considering you have many conference evidences and perhaps innovation might be more relevant? You can move OC4A and other conferences apart from the ones you have mentioned in MC2 to create two strong evidences for OC2. And instead of OC4A conference you can include a letter from your research supervisor.

3

Do you have a patent for any of your work?

4

Your evidence outline shows strong academic credentials with CVPR and NeurIPS papers. The conference talks at SIGGRAPH Asia and ECCV are solid, though you’ll need to strengthen how you present them. For MC1, combining the press coverage with the PI’s attestation letter into one document works well - I’ve seen similar setups succeed when the letter clearly links the applicant to the media coverage.

Your OC3 section needs the most work. The 100 users for your dataset and tracking software is low for this visa level. I’d recommend pivoting to OC2 instead since you have strong conference evidence and review work for NeurIPS and ToG. Move your SIGGRAPH and ECCV talks to OC2 and use your conference organizing or any mentorship activities as the second evidence. For OC4, your three first-author papers at top venues are strong - just ensure your self-documentation explains the impact and citation metrics clearly.

The PyTorch docstring edit is too weak for any criteria. Focus on py-motmetric as your main OSS contribution since you developed it and it has practical applications. Get a reference letter from someone who used your tracker in their research to validate the 6.9mn Euro project impact claim. Your LORs from UCL, Ecole Polytechnique, and INRIA are well-positioned, but make sure each letter addresses different aspects of your work rather than repeating the same achievements.

5

Hey All,

Thanks a lot to everyone who took the time to answer me in detail. Sorry, I couldn’t respond individually, but I incorporated your suggested changes. Thankfully, my stage 1 got approved. :slight_smile:

1 Like