Andrei Ursachi — entrepreneur, neuroscience researcher, author

What I do

Entrepreneur — 17 years building companies; from zero to $50M in combined revenue; led teams of 100+; Forbes 30 Under 30 (2021); Forbes Romania cover (2022); raised €2.5M in funding as founder & CEO of multi-agent AI systems.
Researcher — 1 peer-reviewed paper at Frontiers in Human Neuroscience (March 2026, DOI: 10.3389/fnhum.2026.1781338); public preprints on Research Square spanning consciousness, LLM behavior, and quantum biology.
Intuitive Systems Evaluation — hybrid evaluation of frontier AI (Claude, GPT, Gemini, Grok) combining adversarial testing with trained intuition; detecting model-state shifts, coherence breakdowns, hidden constraints, and deception. Signal before language. Evidence before belief.
Intuitive — 15+ years of contemplative training (Silva Method, the Monroe Institute, among others) — turning inner signal into a decision-making edge.

Intuition, Applied

My work lives where intuition meets rigour. Fifteen-plus years of structured cognitive training have given me a pattern-recognition instinct that cuts across domains.

Trained intuition, scientific inquiry and practical business experience. Different ways into a question, with a clear account of what the evidence supports.

Track record

Adversarial deception detection across SOTA models — Claude, GPT, Gemini, Grok.
Exploratory work on pattern recognition and model behaviour. These explorations are not independently validated performance benchmarks.
Peer-reviewed publication in Frontiers in Human Neuroscience (2026).
12+ research manuscripts across LLM behavior, neuroscience, physics, and quantum biology.
15+ years of structured cognitive training — Silva Method and the Monroe Institute, among many others.
Founder & CEO — built multi-agent AI systems, raised €2.5M in funding; Forbes 30 Under 30 (2021), Forbes Romania cover (2022).

Ventures & Impact

Computational research for difficult decisions

For R&D teams choosing between costly technical directions. I turn existing data into ranked hypotheses, falsification tests and a practical validation plan across drug discovery, materials and selected computational research problems. Direction Preview: €5,000 + VAT where applicable; commissioned computational research from €30,000 + VAT, taking up to one month depending on the mechanism type; the initial fit review is free.

Explore the consulting process

Intuitive Systems Evaluation

A hybrid practice for evaluating complex AI systems — adversarial testing combined with trained intuition. Detecting model-state shifts, coherence breakdowns, hidden constraints, and deception across Claude, GPT, Gemini, and Grok. Intuition generates hypotheses; reproducible experiments decide.

Get in touch

Business services

Presentation websites, online stores, AI-assisted marketing and business automation. Start with the business goal, agree the scope and build a practical solution. Explore the services.

Follow the mechanism. Examine the work.

Explore the papers behind the practice, from neuroscience to mathematical physics. Open each record to read the methods, results and questions still in progress.

Featured research

Research on the Riemann hypothesis · Zenodo · Preprint · 2026-10-25

Local Weil Positivity and Signed Prime-Deficit Criteria

A local positivity certificate and signed arithmetic criteria for research on the Riemann hypothesis, with the remaining all-scale condition stated explicitly.

Zenodo: 10.5281/zenodo.23645233

Yang–Mills programme · O(3) model · Zenodo · Working paper · 2026-10-25

Spectral-gap certificates and signed response for the two-dimensional O(3) model

Finite spectral-gap certificates and signed response in the two-dimensional O(3) model, with explicit premises and open questions for the next stage.

Zenodo: 10.5281/zenodo.23645227 · Research Square · Earlier version: 10.21203/rs.3.rs-11440706/v1

Published research

Frontiers in Human Neuroscience · Peer-reviewed article · 2026-03-04