We build to understand.
We are a research and engineering laboratory investigating how computational systems can help us understand people, skills, and intelligent behavior under uncertainty.
The Continuous Inquiry Loop
Every system we engineer originates with an epistemological inquiry: Can we represent latent competence under uncertainty? Can NLP capture trade expertise without relying on corporate job titles? We build software as an experimental instrument to gather evidence and reveal the next question.
Five Research Principles
We reject corporate agency jargon and 'welcome to our company' platitudes. We build systems to understand how people learn, work, and adapt — and we put the computational hypothesis upfront.
Conventional portfolios reduce engineering to static screenshot cards with vague promotional bullet points. We treat every project as an empirical inquiry with explicit failure modes and reproducible evidence.
Visitors shouldn't just read about algorithms; they should experience the math. Our research pages embed active Bayesian updaters, interactive NLP parsers, and live simulator harnesses.
No generic giant purple AI glows, dashboard clutter, or stock photography. We rely on strict typography, monospace data streams, and disciplined monochrome with a single intentional accent.
We test our models against the real-world friction of vocational accreditation, TVET curricula, and African labour-market signals where documentation is incomplete and noisy.
Grounded in TVET and Developing Labour Markets
Much of contemporary AI research in human capability is tuned exclusively for Silicon Valley software engineering roles. We deliberately ground our work in Technical and Vocational Education and Training (TVET) ecosystems and developing economies.
In these environments, credentials are heterogeneous, occupational frameworks like the South African Organising Framework for Occupations (OFO) are constantly evolving, and informal experience must be translated into verifiable, standardized competencies.