PharmD by training. Close enough to the science and to regulated healthcare to judge what a model is really claiming.
About
Deep science meets shipped systems.
I am Saber Graf, founder of SG AI Solutions. I have spent over a decade at the intersection of complex biological data, software engineering, and machine learning, close enough to the science to judge a model, close enough to the code to ship one.
PhD in neuroscience (Université de Bordeaux, 2024). Over a decade in electrophysiology and machine learning on real biological data.
Engineer by daily practice. I design and maintain an internal cloud-native platform (Django, Next.js, Azure) every day.
How I build
I build and ship on an internal cloud-native platform (Django, Next.js, Azure) that I design and maintain every day: RAG, agents, data pipelines, and machine learning on real biological data. It is why an audit from me comes with sample production-grade code, not slides, and why a build reaches production instead of stalling at the POC.
Peer-reviewed, first-author in Scientific Reports 2025.
Guardrails
What I don't do
No legal AI Act reports. Refer to a lawyer for those. The technical side of the AI Act (logging, monitoring, explainability) is in scope.
No CRO engagements, due to a contractual constraint.
Every engagement starts with an NDA. No client data is used to train models, and I can work entirely within your infrastructure when required.