We investigate technologies before we recommend them. These are the areas we are actively exploring. They are research interests, not published papers or funded programmes.
Each area is framed by the practical questions we want to answer. Findings feed directly into how we design client systems.
R.01
Models chosen and trained for a specific, measurable task, with attention to data quality, evaluation and deployment constraints.
Open questions
R.02
Systems that combine models, rules and human review, so automated decisions remain explainable and correctable.
Open questions
R.03
Integration of perception, language and actuation for robots and interactive devices that work in real environments.
Open questions
R.04
Software-defined radio, signal processing and machine-learning-assisted analysis of the radio spectrum.
Open questions
R.05
Running useful inference on microcontrollers and edge devices, close to the sensor and away from the network.
Open questions
R.06
Evaluating new automation tools and AI agents against real workflows before recommending them to clients.
Open questions
The same method applies to client feasibility studies. You receive the findings, including what did not work.
A specific question and the criteria that would answer it.
The smallest prototype or experiment that can test it.
Results against the criteria, recorded as they happen.
A clear recommendation: proceed, adjust or stop.
A feasibility study gives you evidence before you commit a full budget. Tell us what you want to find out.