Compute and methods, as a partnership
Organisations that hold real data and cannot send it to a commercial cloud service have few places to take a modelling problem. The laboratory is one of them.
Hosted compute and model inference
GPU capacity and open-weight language model inference on Monash hardware. Useful where an AI roadmap is blocked on compute, or where security and regulatory requirements rule out a cloud provider. Data stays under institutional custody throughout.
Analytics migration
Porting existing workflows out of proprietary packages into reproducible R or Python: survival models, risk stratification, forecasting and explainable model output. Delivered with training for the partner's own analysts, so the capability stays in-house rather than returning as a recurring invoice.
Joint publications and grants
Co-authored work in peer-reviewed outlets and joint applications to national and international funding schemes. Partners with data and a mandate, paired with a methods team, tend to be competitive in calls that neither would win alone.
Training and co-supervision
Executive education in quantitative methods, short courses for analyst teams, and co-supervision of higher-degree research candidates working on the partner's problem.
Awaiting confirmation
Awaiting confirmation
Awaiting confirmation
Research partnership
A defined research question, joint design, shared authorship. The partner supplies data and domain knowledge; the laboratory supplies method, compute and analysis. Most suited to organisations with a research mandate of their own.
Consultancy
A scoped piece of analytical work delivered to a specification, with a report and code. Appropriate where the organisation needs an answer rather than a publication.
Compute hosting
Access to laboratory GPU capacity and local model inference for work the partner runs themselves, under an agreement covering data handling, access control and audit.
Talent and training
Co-supervised higher-degree candidates embedded in the partner's problem, short courses for analyst teams, and executive briefings for decision-makers.
Starting a conversation
The most useful first message describes the data (roughly how much, what kind, what restrictions apply), the question, and what has already been tried. That is usually enough for us to say whether the laboratory is the right home for it.