Scientific Oracle · Cross-science hypothesis engine · Private commission

Find the scientific direction worth pursuing next.

I specialize in extracting mechanisms: turning complex observations into clear, testable explanations of how systems work. I combine 15+ years of trained intuition, interoception and pattern recognition with frontier AI to explore unconventional hypotheses. Literature, existing data and computation test what survives. You receive a clear next step, the evidence behind it and a proposal for deeper research when justified.

Bring one consequential question from any scientific field that can be investigated safely through computation. The €5,000 Direction Preview identifies the strongest direction for deeper investigation, or tells you to reframe or stop.

Test the thinking before commissioning the investigation.

One agreed scientific decision is taken through hypothesis formation, a bounded evidence challenge and an accountable recommendation. The Preview shows enough of the route to decide whether deeper research deserves funding. It is not the completed investigation or a final scientific answer. A milestone proposal is included only if deeper work is justified.

Written memo within 7 days · 45-minute prepared working session · one decision

The decision, stated precisely

The one choice the work must move, its constraints and the evidence that would change it.

The strongest surviving direction

The candidate route that best survives the bounded challenge, with enough mechanism and reasoning to understand why it earned attention.

Evidence for and against

A bounded challenge against available literature, public or approved data and computation, including the strongest counterargument.

Assumptions and critical unknowns

What remains uncertain, what is being inferred and what would invalidate the current reading.

Fastest falsifiers

The cheapest useful analyses, comparisons or tests that can discriminate between the live alternatives.

Recommendation and milestone proposal

A pursue, reframe or stop recommendation, plus scoped methods, milestones, timing, fees and stop conditions if deeper research is justified.

Check the fit, or put one question through the method.

Both routes begin with a short, non-confidential brief. The free call helps us decide if the work is a good fit. The paid engagement includes advance research, a prepared working session and a written decision memo.

Private fit call

Free · 20 minutes

A focused conversation about your question, timeline and available evidence. Andrei reviews your summary to assess suitability. Research begins with the separately agreed Direction Preview.

  • Basic, non-confidential intake
  • Fit, timing and boundary discussion
  • No document or source review promised
  • No obligation and no payment

Request the free fit call

Direction Preview

€5,000 + VAT · 45 minutes + memo

After scope acceptance, your brief and a limited source pack are reviewed before the session. We begin with the evidence already on the table. The written memo follows within seven days.

  • Advance brief and limited-source review
  • Prepared 45-minute working session
  • Written pursue, reframe or stop memo
  • Milestone proposal only if justified

Request the prepared Preview

Full computational research begins at €30,000 + VAT and is commissioned separately. Nothing continues automatically, and no confidential files are requested before handling terms are agreed. Full service can take up to 1 month, depending on the mechanism type. The delivery schedule is agreed with the scope before work begins.

Only a direction that survives earns a larger investigation.

If deeper work is justified, the Preview includes a tailored milestone proposal. Full computational research begins at €30,000 + VAT and may include systematic evidence synthesis, public or approved-data analysis, simulation, model comparison, code and adversarial review. Scope, handling, IP, acceptance criteria, fees and stop conditions are negotiated in writing. Nothing continues automatically. Full service can take up to 1 month, depending on the mechanism type. The delivery schedule is agreed with the scope before work begins.

Research & publications

Read the methods, results and open questions in papers spanning neuroscience and mathematical physics.

Peer-reviewed article · Frontiers in Human Neuroscience

Golden ratio organization in human EEG is associated with theta-alpha frequency convergence: a multi-dataset validation study

DOI 10.3389/fnhum.2026.1781338

Preprint · Zenodo

The Riemann Hypothesis and Yang–Mills Theory: Shared Operator Methods and Open Bridges

A comparison of operator estimates for the Riemann hypothesis and lattice SU(2). The preprint reports a certificate for the regularized arithmetic form and a lattice estimate, then identifies the conditions that remain open.

DOI 10.5281/zenodo.24131223

Preprint · Zenodo

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.

DOI 10.5281/zenodo.23596250

Working paper · Zenodo

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.

DOI 10.5281/zenodo.23671651

Publication status is shown for every work. Preprints and research manuscripts are shared for scrutiny and have not been peer reviewed. Featured programmes retain their stated assumptions and open questions.

Read the 2026 paper ↗ · ORCID

Broad scientific scope. A strict hazard boundary.

The practice works with reasoning, literature, code, simulation, models and existing public or specifically approved data. It may help determine whether a physical validation deserves funding. Hazardous execution, direction and advice are outside the scope.

A question may still be suitable when later validation needs a qualified laboratory or specialist. The Preview stops at a defensible computational direction and clearly names where specialist ownership must begin.