Research Projects
Explore our cutting-edge research initiatives and collaboration opportunities.
Join hands with our faculty and students to solve real-world problems.
Guiding Principles for Innovation
We are committed to ethical, impactful, and open research.
A named decision, stated up front
Every project charter opens with the decision it exists to inform and who owns that decision. If neither can be written down, we decline.
Evidence over opinion
Claims are supported by measurement, benchmark or prototype. Where evidence is thin, the report says so rather than dressing confidence over a gap.
A route to implementation
Findings hand over to the engineering floor with a scoped build path — so a positive result can become a working system in the following quarter.
Negative results are results
A well-evidenced "do not build this" is often the highest-value outcome we deliver, and it is reported with the same rigour as a success.
Where we concentrate our efforts
Our labs specialize in several key domains of modern technology.
Machine learning applied to decisions someone is accountable for.
We work on AI where the output feeds a decision with consequences — a clinician’s review, a maintenance call-out, a credit exception. That means evaluation, error analysis and human-in-the-loop design matter more than benchmark scores.
- Can this decision be automated reliably enough to be trusted, and at what error cost?
- Is our data sufficient in volume and quality, or is that the real project?
- Does a large language model outperform a simpler, cheaper approach here?
- How do we evaluate a system whose ground truth is contested?
From Idea to Impact
How a research project progresses from inception to final output.
Scoping & Charter
We convert your question into a research charter: the decision at stake, the hypothesis, the evidence that would settle it, the method, and explicit stop criteria. Signed by you and the research mentor before work begins.
- Research charter
- Hypothesis and method
- Stop criteria and budget
- Data and access requirements
- The decision cannot be named or has already been made
- No available evidence could change the outcome
What is currently on the register.
Placeholder entries. Partner names withheld pending publication consent — replace before launch.
What we produce
Our research results in papers, patents, and open-source software.
Research Report
A rigorous answer to the stated decision, with confidence intervals, methodology, and limitations explicitly documented.
Executive Decision Briefing
A two-page summary for the decision owner, abstracting the method and highlighting the commercial or operational consequence.
Scoped Implementation Path
If the finding supports a build, an architectural proposal, risk assessment and cost estimate to build the system.
Instrumented Prototype
The experimental software or hardware rig built to test the hypothesis, handed over as-is (not production ready).
Dataset or Benchmark
The cleaned, annotated dataset used for evaluation, or the bespoke benchmark developed to test the system.
Partner with Us
Different ways industry and academia can work with our labs.
Sponsored Research
3–12 monthsYou fund the project entirely to answer a question specific to your operation. The research is closed and results are confidential.
Feasibility Study
1–3 monthsA short, bounded investigation to establish whether a larger research or engineering project is actually possible.
Joint Academic Programme
1–3 yearsCo-supervised with university faculty. Longer-term, publishable research tackling a fundamental rather than applied problem.
Grant Co-application
VariesWe partner on government or industry body grant applications where our research capabilities meet the scheme requirements.
Got an idea?
Submit your research proposal for collaboration.
A good proposal states
- The decision waiting on the answer, and who owns it
- What you already believe, and how confident you are
- What evidence would change your mind
- Any data, systems or site access you can provide
Joint academic programmes
We co-supervise with university faculty, host student researchers on site, and support joint publication and grant applications. MoUs are handled by the programme office.
The cheapest research is the kind that stops a bad build.
Tell us the decision. We will tell you honestly whether it can be answered, what it would cost, and whether we are the right people to answer it.