Our paper, “VAMAE: Vessel-Aware Masked Autoencoders for OCT-Angiography,” has been accepted to the 28th International Conference on Pattern Recognition (ICPR 2026).
This work sits around a problem I care deeply about: how to make medical vision models learn better when labeled data is limited.
In medical imaging, labels are not cheap. They depend on expertise, time, careful annotation, and access to the right clinical context. This is especially true in specialized imaging domains like OCT-Angiography, where the structures of interest are subtle, complex, and clinically meaningful.
So it is not enough to build models that only perform well when large labeled datasets are available. We need methods that can learn useful representations from the data itself, while paying attention to the structures that actually matter.
With VAMAE, we explored a more structure-aware approach to self-supervised learning for OCT-Angiography retinal imaging. Instead of treating all image regions the same during masked pretraining, we guide the model toward vessel-rich regions, encouraging it to learn from the anatomical structures most relevant for downstream retinal analysis.
That idea is simple, but important to me:
Medical AI should not just learn from pixels. It should learn with some respect for the structure and meaning inside the image.
I’m thankful to have worked on this with @llerioluwakiiye Abolade, Prince Mireku, Kelechi Chibundu, Peace Ododo, Promise Omoigui, and Solomon Odelola.
Prince Mireku will be presenting the work on behalf of the group at ICPR 2026 in Lyon, France.
Paper: https://t.co/95vUhvb4ZL
Many thanks to the @ml_collective community for creating spaces that support and inspire early-career researchers.
Excited to be hosting a conversation with @useaccrue this Saturday on:
How to Start Building Wealth in Your 20s 💸
Register here to join us: https://t.co/fW494LxF1F
Databricks Student Fellows Program - Exclusive opportunity for students passionate about data, AI & computer science
Benefits include hands-on platform training, free/discounted industry certifications & pathways to internships at Databricks and partner companies
https://t.co/A3Wd9FhOx1
In April, I attended a courtesy visit paid by the leadership of the Nigeria Computer Society (NCS), Lagos Chapter, to the University of Lagos @UnilagNigeria to discuss collaborative pathways in shaping the next generation of technology professionals.
I borrowed an umbrella from my Airbnb host in Kyoto. I forgot to return it when I checked out, and realized when I was already on the train to Osaka.
I felt terrible. It was a nice umbrella, not a cheap one. I messaged the host apologizing.
She responded: "No problem! Enjoy the umbrella. It's yours now."
I said I'd mail it back. She said "please don't. Postage costs more than an umbrella. Just use it and think of Kyoto when it rains."
I insisted I wanted to return it. She said "okay, but I have a different idea. Next time you see someone who needs an umbrella and doesn't have one, give them this umbrella. Tell them to do the same when they are finished with it. Maybe an umbrella travels all around Japan helping people."
That idea was so beautiful I agreed.
Two weeks later I was in Hiroshima and it started pouring. A woman with a baby was standing under an awning looking stressed. No umbrella, the baby was crying.
I walked over and gave her the umbrella. Told her the story in broken Japanese. She understood enough.
She tried to refuse but I insisted. Told her "when you're done with it, give it to someone else who needs it."
She nodded, said thank you about ten times, and hurried off with her baby.
I got soaked walking back to my hotel but felt good about it.
Sometimes I wonder where that umbrella is now. Hope it's still traveling, still helping people.
I recently supported and mentored a team of students in the Afretec UNILAG Inclusion Team Innovation Challenge, together with other mentors.
It has become one of the most meaningful projects I have worked on.
The challenge focused on assistive technology in higher education: how we can make learning more accessible for students with disabilities from the very beginning, not as an afterthought.
The team came with the idea and lived experience.
I helped shape the product direction and built the application to turn their concept into a working solution.
The result was AccessLearn: an accessible learning platform that helps students engage with educational materials in formats that work better for them.
It includes:
• Text-to-speech in English, major Nigerian languages, and Pidgin
• AI-powered translation
• A fingerspelling tool
• Accessibility profiles
• An AI tutor for document-based questions
The team went on to place 2nd in the competition.
I am proud of them not just because they placed, but because they worked on something with real human value.
AccessLearn reminded me that accessibility is not a feature to add after a product is built.
It is a design responsibility from the beginning.
Grateful to the University of Lagos, Afretec Network Inclusion Team, Theseabilities Foundation, the organizers, and the other mentors who created a space where students with disabilities are not only included, but empowered to innovate, lead, and build.
Ambitious people are my type. I can’t even hide it, once you start talking about doing impossible work, the kind of talk that inspires greatness, you have my allegiance till the wheels fall off.
We may fail (woefully) but we will never be guilty of not trying.
🚀 Today, I'm officially launching @ArushaWealth , a registered and regulated financial planning & wealth management firm for immigrants and children of immigrants building their American Dream.
Over the past 18 months, I've directly helped 300+ people get their green cards & indirectly helped so many more. But there’s one reality we keep seeing: that's just the beginning of the journey, not the end.
We arrive with ambition & do everything right: we get into top schools, earn valuable degrees in the right fields, build strong careers & earn well, only to realize over time that high income doesn’t always translate into high net worth or lasting wealth.
You're earning well but you see how much goes to taxes
You're not sure how much should go to paying off your past student loans and how much to investing for your future
You have family obligations back home that you can't ignore & you're not sure how to manage this without sacrificing your own goals
Without a clear long term plan, you wake up down the line & realise you are stuck - permanently dependent on your paycheck to maintain the lifestyle you have.
This is not the American Dream we came here for.
We have officially launched @ArushaWealth a financial planning & wealth management firm for immigrants and children of immigrants in the United States. Built on my experience as a wealth advisor at Goldman Sachs Family Office, along with my training as a Certified Financial Planner® and my MBA from Columbia University, Arusha helps individuals & families put structure around what comes next through comprehensive financial planning, tax planning, investment management, and legacy and estate planning.
I've worked with some of the wealthiest American families and seen exactly how wealth is built, protected & passed down in the U.S. With Arusha, our focus is on bringing to immigrant families that same level of planning and the opportunity to build lasting prosperity.
Our clients so far include startup founders who have raised over $10M, and professionals across tech, finance, and law, including Microsoft, Amazon, the World Bank, private equity, and Big Law.
👉 If this resonates, feel free to reach out here: https://t.co/BzM2uwAYS3
And if you’d like to stay in touch & receive occasional updates and guidance, you can subscribe to our newsletter here: https://t.co/jUwsCu9LGE
This is just the beginning. I’m excited to work with the next generation determined to build lasting wealth.
Basic things that have turned luxury in this country
- Light
- Clean water
- Turkey
- Egg
- Good road
- Education
- Light
- Healthcare
- Traveling via Air
- Traveling by Road
- Light
- Owning a car
- Light