About me

I’m Nassim, a Research Engineer working on machine learning for ophthalmology.

I built and operate OphtaFlow AI at Quinze-Vingts Hospital in Paris: diagnostic support running in production, available 24/7 and in beta testing with physicians. CorneaForge is the on-premise engine behind it, connecting device ingestion, scientific image processing, automatic prediction and research.

I work across the whole path: extracting measurements from device exports, implementing corneal geometry, and bringing images and diagnostic predictions into the physician's examination workflow. I also train image models and build the research systems that let me test new approaches.

Clinical AI in operation

CorneaForge’s deployed services ingest MS-39 and Corvis data, process OCT images, compute features, and render examination views. MinIO retains source objects, PostgreSQL holds structured measurements and processing state, and a FastAPI application serves the OphtaFlow AI interface. Dedicated workers run under systemd on the hospital’s infrastructure.

Physicians can inspect examinations, review diagnostic predictions alongside the source measurements, view native videos, and annotate findings in one application. Research dataset builders use the same source lineage while defining the population and variables for each study. My architecture article follows these paths through the deployment and explains why I separated storage, background processing, and interactive access.

The prediction service works continuously: once an examination has the required extracted features, it calculates missing results and saves them. Diagnostic results are available in under one minute after examination completion. Entering the patient's identifier retrieves those completed calculations for the physician.

Where I am taking OphtaFlow AI

My January 2027 targets are diagnostic support for 100+ diseases, fusion of five corneal and retinal devices, and more than one million examinations processed. I am prioritizing self-supervised learning, starting with JEPA, as the foundation for that expansion.

Diagnosis is one task in the broader project. My roadmap also covers treatment and surgery recommendations, calculated and prefilled surgical parameters, and estimates of success and their uncertainty. The project targets connect this clinical direction to the research and production systems I am building.

Experiments and engineering results

For anterior-segment OCT, I implemented five pretraining recipes on a shared index of 3,427,130 B-scans: full-field masked reconstruction, two matched patch-budget variants, diffusion reconstruction, and global–local JEPA. The recorded results show distinct representation geometries: the patch-limited arms concentrate variation, while JEPA spreads it across more directions. I explain the measurements and their interpretation in Reading the geometry of AS-OCT pretraining.

I also implemented and optimized JEPA across two NVIDIA L40S GPUs. Profiling traced a communication stall back to late input preparation. Overlapping that work reduced update time from 10.246 to 8.787 seconds in a matched comparison, with input, gradient, optimizer-state, and resume checks preserving the experiment.

My Corvis research combines classical corneal region segmentation, motion analysis, and an experimental neural video classifier. I built synchronized views and review tools that show the extracted corneal regions, their motion, and the classifier’s output.

ML Training Monitor grew out of this work: a public monitoring and profiling workflow for coding agents. OCT-CUDA explores kernel fusion for OCT reconstruction with cuFFTDx, focusing on the cost of moving intermediate values through GPU memory.

From statistics to research engineering

I came to machine learning through econometrics: a bachelor’s degree in Economics at Paris-Est Créteil, followed by a master’s in Econometrics and Statistics at the University of Angers. I joined Quinze-Vingts in April 2024, initially as an intern, and continued as a Machine Learning Engineer.

That background connects statistical modeling with the systems I build today. I collaborate with clinicians on the measurements and outcomes that define a useful research question. My publications cover surgical outcome prediction and corneal biomechanics; my articles explain the computational work behind the research.

Get in touch

If you’re working on ophthalmic ML, scientific computing, or the systems behind research, I’d be happy to compare notes. You can reach me at louissi.nassim@gmail.com.