Hi, I'm Abdul.

This will be my Company's website replacement, for now. At a minimum, I hope it provides more useful information.

If you're reading this, it's likely I reached out to discuss a problem that isn't unique to you, but common across the industry. This note is simply meant to provide a bit more context - non-essential, but relevant. As I try to keep emails brief and focused, this page exists to go deeper into the problem and who I am. Again, not at all essential, but simply just clarity on the type of person who you (or your group) could potentially be working with :)

I am Abdul. I studied Math at the University of Waterloo and the Fields Institute, and am currently on leave from Dartmouth University. In my undergraduate I realized I particularly enjoy Machine Learning and Statistics from an Abstract Perspective - i.e Topological Data Analysis and Cohomology being some of my favorite courses.

I also spent time in my undergraduate researching various niches of gradient descent algorithms, and worked on optimizations to such algorithms to improve convergence in image classification tasks. Admittedly, while I derived novel bounds, my additions were incremental improvements to the frontier rather than world-changing. Regardless, I mention this experience as it only made me more certain that If economics were divorced from reality, I'd likely be working on becoming a researcher.

My current bet is that as AI labs scale energy, compute, and infrastructure, healthcare will remain a disproportionally lagging domain - simply due to massive data bottlenecks. This is not a problem architectural or algorithmic changes can solve, because verifiability remains the core constraint.

There is no broadly effective solution to these data constraints today. As a result, the vast majority of researchers are forced to work around them in one or more of the following ways:

These approaches are not solutions to the underlying problem; they are coping mechanisms that narrow the region of questions academia can realistically pursue. Large parts, if not most, of healthcare research is simply inaccessible with unannotated data - especially work that depends on explanation, physician reasoning, attribution, rare events, and verifiable outcomes.

It becomes clear that the current approach will not enable progress in healthcare at an exponential scale. This is the problem that I intend to solve.