Federated AI for Vision Science & Neuroscience · Pre-Seed 2026

Federated Analysis for
Vision Science and
Neuroscience Research

UndosaTech helps vision science and neuroscience researchers run secure, governed, federated analysis across institutional datasets — without moving raw patient data. Built with NHS Trusts and research universities, designed for the regulatory reality of UK and EU health data.

$42B
Annual global neuroscience R&D spend
95%
CNS drug trial failure rate at Phase III
12+ mo
Average cross-institution data access wait
80%
Researcher time lost to data wrangling


JO
Dr. John Ohanebo
MBBS · MSc Applied Neuroscience, University of Dundee
Member, British Neuroscience Association
Founder & CEO, UndosaTech
Neuroscience
Vision Science
Substance P
Myopia Research
Biomedical AI
NHS

Why We Exist
Born From a Researcher's Frustration

"During my research into the role of Substance P in the axial elongation of myopia, I hit a wall. The data I needed was scattered across institutions, locked behind governance barriers, and months away from access. I was spending more time fighting for data than doing science."

UndosaTech was built to solve exactly that problem — starting with the vision science and neuroscience research community. We are building the infrastructure that has been missing — secure, federated, governed — one research community at a time.

The Problem
Vision Science and Neuroscience Research Is Sitting on Data It Cannot Connect

A researcher may have access to data at their own institution — but the datasets they need sit at hospitals across the country, locked behind governance requests, or buried in lab servers as unpublished dark data that science has never seen.

$160B
Wasted annually in global medical research
Over 80% of medical research funding is wasted — largely because dark data is never reused or built upon.
50%+
Clinical trial results never published
Over half of clinical trial results remain unpublished 30 months after completion. The FDA is now cracking down.
12+ mo
Average data access wait time
NHS Information Governance requests can take over a year to resolve — if approved at all.
95%
CNS drug trial failure rate
Driven by poor patient stratification from inadequate access to cross-institutional data at earlier research stages.
"Existing medical data is not fully exploited by machine learning primarily because it sits in data silos and privacy concerns restrict access. Without access to sufficient data, ML will be prevented from reaching its full potential and from making the transition from research to clinical practice." — Rieke et al., The Future of Digital Health with Federated Learning, npj Digital Medicine, 2020.

Responsible Research Data Reuse
Most Research Data Is Used Once, Then Archived. Science Pays the Price.

Published papers represent a fraction of the data institutions generate. Negative results, inconclusive trials, and archived cohorts contain extraordinary scientific value — but no current mechanism exists for institutions to make this data available for governed secondary research, recover the cost of stewarding it, or share in the public benefit it produces.

UndosaTech's Research Data Reuse Programme is designed to change that — within the framework of institutional governance, patient consent, and ethics committee oversight. Institutions retain full control. Data never leaves its source. Reuse happens only under explicit governance agreements, with full audit trails and clear public-benefit criteria.

PUBLISHED DATA
What the world can see

Summary statistics. Effect sizes. P-values. Conclusions drawn from data that the reader can never access. Science built on shadows of the underlying reality.

UNPUBLISHED & ARCHIVED DATA
The data that informs better science

The raw OCT scans. The MRI sequences. The negative results. The inconclusive trials. The unpublished cohorts. Under proper governance, this data can support more representative, more reproducible science — informing models that current published datasets alone cannot.

2× more
Positive results published vs negative
Publication bias means much of the most informative data — including what didn't work — is hidden from science.

Cost recovery
A governance-first model for stewardship
Institutions can sustain the cost of long-term data stewardship through transparent reuse arrangements — with full governance and audit trails built in.

FDA
Regulatory direction in the US and EU
In April 2026, the FDA contacted 2,200+ sponsors about unreported trial data. Across the US and EU, regulators are pushing toward greater secondary use of research data under stronger governance frameworks.

Our Solution
The Intelligence Travels to the Data

A cloud-native SaaS platform unifying fragmented biomedical datasets with no-code AI tools — within HIPAA- and GDPR-aligned frameworks. We do not ask institutions to hand over their data.

🔒
Zero-Trust Federated Learning

AI models travel to local institutional servers to train. Datasets never leave the institution. This architecture is designed to support institutions in meeting GDPR, HIPAA, and EHDS obligations — reducing the data-transfer burden while keeping institutional governance, consent, and audit responsibilities intact.

🧠
No-Code AI for Domain Experts

Automated model building, predictive analytics, computer vision, and deep learning workflows — so neuroscientists and clinicians stay in charge of the science without needing a data science team.

⚖️
Enterprise Regulatory Compliance

Built around HIPAA, GDPR, and the EU Health Data Space Secure Processing Environment. Automated data lineage, audit trails, and support for Predetermined Change Control Plans where FDA submissions are in scope.

🌑
Research Data Reuse Programme

A governed framework that enables institutions to make archived and unpublished research data available for ethical secondary analysis — under their own governance, with patient consent honoured, and with public-benefit criteria built in. Institutions are supported in recovering stewardship costs through transparent, audit-friendly arrangements.

🔬

Rooted in Real Research: UndosaTech emerged directly from active neuroscience research — specifically a literature review into the role of Substance P in the axial elongation of myopia. The data fragmentation encountered during that process is the exact problem this platform is being built to solve. Research areas supported include Alzheimer's & dementia, Parkinson's disease, neurodegeneration, vision science & retinal imaging, treatment-resistant depression, and chronic pain.

Competitive Landscape
A Specialised Wedge in a Crowded Space

The federated learning and biomedical data infrastructure space includes well-funded players such as Owkin, Lifebit, Rhino Federated Computing, and TriNetX, alongside national health data initiatives across the UK and EU. UndosaTech does not aim to compete with these platforms as a general-purpose biomedical infrastructure provider.

Instead, UndosaTech focuses on a specific, underserved wedge: vision science and neuroscience research, with a governance-first reuse framework, an institutional cost-recovery model designed for NHS Trusts and research universities, and a no-code interface built for domain researchers rather than data engineers.

Specialisation, not breadth, is the strategy.

Business Model
Five Revenue Streams Built for Network Effects

Predictable recurring SaaS income combined with marketplace and partnership revenue that scales as the network grows.

01 — SAAS
Platform Subscriptions

Annual licences tiered by institution size: Starter (small labs), Professional (departments), Enterprise (pharma / NHS Trusts).

02 — MARKETPLACE
Transaction Fees

A percentage of each royalty transaction on the Neuro-Marketplace when anonymised datasets are licensed to third parties. Scales with network size.

03 — PARTNERSHIPS
Pharma Data Agreements

Bespoke data access agreements with pharmaceutical companies for curated, federated neuroscience datasets. High-value, low-volume contracts.

04 — COMPLIANCE
Compliance-as-a-Service

Premium tier covering automated regulatory reporting, data lineage tracing, and PCCP management for FDA submissions.

05 — API
API Access

Usage-based pricing for organisations integrating UndosaTech's AI models and data connectors into their existing research pipelines.

YEAR 2 TARGETS
Unit Economics

Enterprise ACV: £85K  ·  Academic ACV: £18K
Gross Margin: 72%  ·  CAC Payback: <14 mo
Net Rev Retention: 115%+

The Team
Founded by a Clinician-Scientist. Building the Best.

UndosaTech is founded by a medical doctor with a neuroscience research background. We are actively building our founding scientific and engineering team.

JO
Dr. John Ohanebo
Founder & CEO

MBBS · MSc Applied Neuroscience (University of Dundee) · Member, British Neuroscience Association. Founded UndosaTech from direct experience of the data fragmentation problem in neuroscience research.

ML
ML Engineer — Hiring
Federated Learning Specialist

Seeking an experienced ML engineer with background in privacy-preserving machine learning, federated systems, or biomedical AI. Equity available. Remote-friendly.

SA
Scientific Advisor — Seeking
Neuroscience / Clinical Research

Actively seeking a Scientific Co-Founder or Advisor from a leading NHS Trust or research university with expertise in neuroscience, biomedical data, or clinical AI. Equity available.

Open to Scientific Co-Founders and Advisors. If you are a neuroscientist, clinician-researcher, or biomedical data expert at an NHS Trust or research university who has felt this problem firsthand — we would genuinely value a conversation. Get in touch →

Why Now
Three Macro Forces Are Converging

The timing for UndosaTech is not arbitrary — three structural shifts are creating a narrow window for the right infrastructure to dominate this space.

01
The EU Health Data Space Goes Live 2025–2026

Cross-border secondary use of health data is now legally enabled for the first time across the EU. Institutions and platforms with compliant infrastructure in place first will capture the market. UndosaTech is designed natively for this moment.

02
Foundation Models Are Commoditised — Data Is the Moat

GPT-4 and its successors have democratised model capability. The competitive advantage in biomedical AI now lies entirely in proprietary, high-quality, labelled datasets. UndosaTech sits on that asset — and builds the infrastructure others depend on to access it.

03
NHS Institutions Are Under Acute Financial Pressure

The royalty marketplace directly addresses a real and urgent need — turning archived patient cohorts into a revenue stream resonates immediately with budget-constrained Trusts and universities. The incentive to participate is financial, not just altruistic.

Get Involved
Investors. Researchers. Partners. Advisors.

We are raising £900K pre-seed and building our founding team. If you understand the intersection of healthcare, AI, and regulatory complexity — we want to hear from you.

Founder
Dr. John Ohanebo
MBBS · MSc Applied Neuroscience
Pre-Seed Round
£900K target
SAFE / Convertible Note · 18 months runway
Location
Dundee, Scotland, UK
Building for the NHS & global research market

UndosaTech
The Data-Innovation Bridge

© 2026 UndosaTech Ltd · Dundee, Scotland · contact@undosatech.com
Confidential · Pre-Seed Stage