Aitele Research — Who we work withEdition 01 — 2026
§ For academics & researchers

Academics and researchers: take your research past the paper.

Most research ends in a paper. At Aitele Research it ends in working technology. Propose a collaboration with one of the labs, apply as a fellow with research of your own, publish through the commons' peer-reviewed Journal of the Commons, and build what the research makes possible, under a compact that keeps at least 60% of it yours.

§ I — What you can do with Aitele

Five ways in.

01

Propose a collaboration

Propose a joint research direction with the commons: your institution, the lab or problem, and what working together would look like. It goes straight to a steward, who replies personally.

Research collaboration →
02

Become a fellow

Apply with your original R&D: research, a method or a working prototype. Fellows are selected on the problem they want to work on, not their pedigree, and fellows read every application and vote.

Apply with your research →
03

Publish in the open

The Journal of the Commons is the commons' own peer-reviewed journal. Manuscripts submitted through the commons appear once accepted as preprints, and carry their live status through peer review and translation.

The journal ↗
04

Turn research into technology

A fellow's research can grow into releases and ventures. The compact covers everything a fellow makes, from papers, code, models and designs to the ventures that grow from them, and the fellow keeps at least 60%.

How the commons works →
05

Take on an open problem

Every lab publishes the open problems it is working on, listed below. Bring one you can move.

The open problems ↓
§ II — How it works

How fellowship works.

  1. 01

    Apply

    One short form about your research. You get an email at once, with a private link to follow your application.

  2. 02

    Read

    Fellows read it and vote. A steward may write with questions, and you can answer from your link.

  3. 03

    A conversation

    If you're shortlisted: one working conversation about what you've built. Not an interview.

  4. 04

    Decision

    Admitted: your Fellow account, and your research set up as your venture. You keep at least 60% of it.

§ III — Labs and ventures

Every lab, and its open problems.

The problems each lab is working on now, and what the commons has published. Live from the commons' own record.

Labs · 10

All labs →

AI & Machine Learning

What if the data built the network?

Open problems
  • Zero-training synthesis of transformer architectures from observed data
  • Reverse Synthetic Neural Networks for resource-constrained and edge deployment
  • Closed-form alternatives to iterative optimization in deep learning
Visit the lab →

Defense Systems

AI that holds when the link doesn't.

Open problems
  • Field-adaptive recognition that re-tunes to a new theatre in seconds, training-free, on edge hardware
  • On-device CDR / tower-dump network-structure typing — honest about the limit that structure is not intent
  • Cadence-biometric access control with no cloud and no enrolled secret to steal
  • GPS-denied and link-denied operation: intelligence that survives a contested electromagnetic environment
Visit the lab →

Stochastic Computing

Computing with probability as the primitive.

Open problems
  • Noise-native inference on edge hardware
  • Stochastic primitives for probabilistic programming
  • Energy-proportional computing for always-on sensing
Visit the lab →

Market Computation

Markets as computational substrates.

Open problems
  • Universal computation via orderbook microstructure dynamics
  • Adversarial amplification and the benefit of selfish agents
  • Non-custodial portfolio construction through market-native algorithms
Visit the lab →

Blockchain Technology

Proof of origin that survives transformation.

Open problems
  • Transformation-invariant fingerprints — provenance that survives paraphrase, re-encoding, translation, and AI remixing
  • Zero-knowledge proofs of priority and derivation that never disclose the protected work
  • On-chain verifiable similarity — proving a later work is, or is not, derived from a registered original
Visit the lab →

Astronomy

A uniform standard of evidence for other worlds.

Open problems
  • Amortized neural posterior estimation for atmospheric retrieval at archive scale
  • Calibrated abiotic-null tests for contested biosignature claims (DMS, CH₄/CO₂ disequilibrium)
  • Stellar-contamination and reduction-systematics modeling for M-dwarf transmission spectra
Visit the lab →

Intelligent Machines

The brain for the next robot.

Open problems
  • Pre-synthesized RSN neuron libraries that compile a control model for each robotic capability — hand/grasp articulation, bipedal gait and balance, gaze and mouth/facial actuation — directly from human motion capture, without gradient training
  • Closed-form motor-primitive synthesis: human-like movement learned from a handful of demonstrations, with no reinforcement-learning rollouts
  • Real-time, on-device inference for actuator control on low-power robot hardware — no datacentre in the loop
  • A portable capability marketplace: drop-in skill models (walk, grasp, speak) that transfer across robot chassis
Visit the lab →

Computational Biology

From sequence to therapy.

Open problems
  • Generalizable models of cellular perturbation response
  • Interpretable structure prediction for intrinsically disordered proteins
  • Low-data regimes in rare-disease therapeutics
Visit the lab →

Quantum Computation

Useful computation before fault tolerance.

Open problems
  • Practical benchmarks for near-term quantum advantage
  • Hybrid classical-quantum optimization at scale
  • Error mitigation without full error correction
Visit the lab →

Tantra Vidyā

The classical sciences of sound, breath and time — made computable.

Open problems
  • Primary-source reconstruction — rebuilding Tantric, Svarodaya and Jyotiṣa tables from scanned editions, with each text's own counts asserted by machine
  • A computable Swara Code: minute-level breath (svara) and element (tattva) expectations from chart and sky, and pre-registered trials against observed breath
  • Listening instruments for mantra practice — on-device syllable, phoneme and in-ear signals, every reading graded by its evidence
  • Calendrical astronomy at reference precision: eclipses, tithis and the Pañcāṅga agreeing with NASA, USNO and Drik to the minute
Visit the lab →

Research and the journal

15 papers, preprints and notes on the research page, and 12 articles in the Journal of the Commons.

§ IV — Questions

Asked often.

How do I propose a research collaboration?
Use the research collaboration form on the contact page: your institution, the lab or problem you're interested in, and what the collaboration would look like. It goes straight to a steward, who replies personally.
Who can become a fellow?
Fellows are selected through an application that looks at the problem you want to work on, not your pedigree. Apply with your research, a method or a working prototype.
Where is the research published?
On the research page and in the Journal of the Commons, the commons' own peer-reviewed journal. Manuscripts submitted through the commons appear once accepted as preprints, and carry their live status as they move through peer review and translation.
Who owns the research I do as a fellow?
It's co-owned under the compact: you keep at least 60%, and the commons' share, at most 40%, goes into the pool. That covers everything you make as a fellow: papers, code, models, designs, and the ventures that grow from them.
Which fields do the labs cover?
AI and machine learning, intelligent machines, defence systems, stochastic computing, market computation, blockchain, astronomy, computational biology, quantum computation and Tantra Vidyā. Each lab's page lists its open problems.