A peer-reviewed foundation, and a staged plan to test the physics of intelligence.
HeyDonto Labs
HeyDonto Labs is a research laboratory. The work is deep learning, quantum-inspired methods and the geometry of learning, applied to data. The lab is built around Dr. Reza Nehzati, the scientist behind the law of natural intelligence, published as Data Field Theory.
Eight papers are in print across Frontiers, Elsevier and Tech Science Press.
Berlin · Mexico City · Knoxville
At fourteen, Reza Nehzati took Iran's national college entry exam, a test written for sixteen and seventeen year olds, and posted the twelfth highest score in the country. He entered college early. He holds a PhD in AI and machine learning. The work since has run through deep learning optimization, quantum-inspired computation and the geometry of learning. He is first author on every paper the lab has indexed, sole author on all but one, and sole named inventor on every patent it has filed. Axiomera and Conduit both run in production on the harmonization pipeline he architected.
His Data Field Theory framework was published in Frontiers in Big Data in June 2026, handled by an editor at Lawrence Berkeley National Laboratory and reviewers at the University of Massachusetts Medical School and Université Ibn Zohr.
The evolutionary neural networks application, USPTO 19/181,522, allowed April 2026 with the issue fee paid.
His federated learning study predicted 30-day hospital readmissions across 47 US healthcare institutions, published in Elsevier's Informatics in Medicine Unlocked.
His record outside the lab includes a wearable graphene-patterned gas sensor in Results in Optics, a graphene terahertz electronic nose in Sensors and Actuators A: Physical and a wireless semi-implantable brain-recording microsystem in Next Research. ORCID · Google Scholar.
Sits on the American Dental Association committee drafting the AI Model Fact Label standard.
Data Field Theory, a physics-based, geometric account of learning, and the foundational model being built on it.
Systems that repair and restructure themselves, from quantum-inspired self-healing code generation to evolutionary optimization of deep learning.
Deep learning models trained jointly across institutions that cannot share raw data, without moving a record between them.
The semantic intelligence and FHIR-native harmonization work, from a quantum-inspired, many-worlds harmonization framework now under peer review to the Epic Clarity study across three health systems, 127,834 patients and 492,542 encounters. Axiomera and Conduit run in production on this research area.
Data Field Theory treats learning as geometry on curved space rather than as a pipeline problem.
A foundational model built on the framework is in development at DFT Labs, announced in July 2026 as a staged plan to test the physics of intelligence.
Axiomera standardizes fragmented healthcare data to the standard the customer already works in, binds it to shared clinical terminologies, and harmonizes it inside the customer's own cloud. Nothing has to be copied out to a vendor environment first, which is usually the reason these projects stall. The analysis and evidence layer sits on top of the harmonized data, and the same pipeline serves pharma, payers, health systems and life sciences.
axiomera.com
Conduit moves referrals, records and clearances between medical and dental providers, on whatever software each office already runs. A physician refers a patient to a dentist and gets the clearance back. A dentist refers out and the records follow the patient instead of arriving as a fax at the other end. The exchange is FHIR and HL7 native, and Conduit is a member of the Oral Health Interoperability Alliance.
conduitdental.ai
The multi-agent development platform the lab built in-house when Cursor, Replit, Lovable and the rest did not do what we needed. Describe a change in chat and it comes back as a reviewed, tested pull request on GitHub, moved there by a team of agents through six stages that fail closed rather than by a single bot that writes code and hopes. Every venture here runs on it, and Red Wheelbarrow is built with Red Wheelbarrow.
redwheelbarrow.ai
DFT Labs is where the geometry becomes a program. Data Field Theory treats intelligence as a field on curved space rather than as a stack of heuristics, published in Frontiers in Big Data in June 2026: one equation, four cardinal predictions, and the disagreements with the current consensus documented in public. A foundational model built on the framework is in development, announced in July 2026 as a staged plan.
dft-labs.comFounder of DFT Labs. His self-healing code generation and self-evolving AI frameworks are published in Frontiers in Artificial Intelligence, and a book-length treatment of Data Field Theory is in preparation.
Founded and sold three software companies, CloudRetriever, Vet Marketing Pro and Veterinary Dashboards, negotiating each sale himself. More than fifteen years building and scaling technology companies. Studied at UC San Diego.
Led the technology diligence for WhiskerCloud's acquisition by Petvisor as its SVP of Technology. Now runs operations for the lab and its ventures.
Built WCG's clinical-trials book from zero to more than twenty million dollars in three years and closed six Fortune 500 clients along the way. Before that, GYN surgical sales at Hologic and nearly two decades of commercial work in clinical research and medical devices.
If you want to see what sits behind the papers, or you work on the same problems from a different angle, book a technical briefing with the people who did the research.