Research Directions.
The laboratory is young; its founder’s experience is not. These are the directions where we take on research engagements today. As client projects complete, this page will show cases instead.
Forecasting & resource optimisation
Demand, load, or staffing behaves in ways the current model cannot explain; the cost of over- and under-provisioning is known and painful.
The driving factors are unidentified — no amount of implementation fixes a model built on the wrong mechanism.
A validated forecasting approach with measured improvement over your baseline.
Applied LLMs & generative models
An LLM demo impressed everyone; production quality, cost, or reliability did not.
Prompt-level fixes are guesses until generation quality is measurable — the uncertainty is in the evaluation, not the API.
An evaluation framework and a tuned pipeline whose quality is demonstrated, not asserted.
Process mining
The process map says one thing; throughput, costs, and complaints say another.
The real process is hidden in event logs and nobody knows its actual shape.
The discovered process, its measured bottlenecks, and quantified improvement scenarios.
* Directions reflect the founder’s 8+ years of industrial research and data-science practice. Past employment engagements are the founder’s professional experience, not this company’s project history — we prefer that distinction stated plainly.