The model is one part. The system is the product.
From the network to the model to the regulation on top: complete systems for energy, industry and health, built to be audited.
Assistants that cite the clause. Agents that know their limits.
Retrieval that answers with the governing paragraph, agents behind typed permissions, and open models on your own servers when the data cannot leave the building. Each one measured before it ships.
What I build with AI →Engineering where failure has a price.
Segmented industrial networks, medical software under MDR, sustained 1 kHz acquisition: the link between plant and control is deliberate, guarded and logged.
How I work →From raw data to a decision that defends itself.
Constrained optimisation, eligibility scoring, event execution under 100 ms on public blockchains, ERP cutovers rehearsed with a rollback: capabilities that dock around one core.
Capabilities & stack →Four systems, built and run solo.
Prediction markets, public funding, weekly menus and tender analysis: the same architecture, shipped four times, documented like engineering.
See the projects →If the hard part is the system, let's talk.
Tell me what you are trying to do and what constrains it. If a model is not the answer, I will say so in the first conversation.