Impact that compounds.
Evidence over claims.
CentPol converts funding into three durable things: trained, publicly-recognized talent; open artifacts institutions can reuse; and AI deployments that are evaluated and governed rather than guessed at. This page is the honest ledger of that conversion.
From policy intelligence to evaluated deployment
The chain is deliberate: judgment first, talent second, artifacts third, deployment last. Skipping steps is how institutions end up with slideware AI. Funding any link strengthens the whole chain.
Proposals, research, structured courses, and cohort sprints build the judgment institutions need before they deploy anything.
Sprints and fellowships turn learners into contributors — every cohort ships a memo, a pitch, an evaluation, or a concept note, and completion is publicly recognized.
Benchmarks, model cards, evaluation reports, governance patterns — published openly in the Evaluation Commons so every engagement leaves a reusable trace.
Institutions in our region choose and run AI systems on evidence — with local benchmarks, trained talent, and governance built in.
The base is live, not hypothetical
Everything below is running now and inspectable on this site — the discipline of showing real state is part of how an evaluation institution earns trust.
Completed sprint cohorts
Cohorts that finished real work — decision-ready memos and pitches produced against real completion standards.
The Evaluation Commons
Benchmarks, model cards, evaluation reports, and governance patterns, published openly as they ship.
Published proposals
National-scale frameworks led by the Uganda AI workforce thesis — the policy intelligence the lab is built on.
Live courses & sprints
Structured curricula and cohort sprints with real completion standards and verifiable certificates.
A verified community
LinkedIn-verified membership, public member badges, and a growing contributor network.
Verification, not vanity
Certificates carry IDs anyone can check. Recognition is published by consent after completion. Artifact counts on this page are computed live from the database — never hand-edited.
Why a shilling here outlasts the engagement
CentPol is structured so that money in produces public goods out — not billable hours that evaporate.
Every engagement leaves a reusable trace
A funded cohort produces trained fellows AND their published deliverables. A funded evaluation produces a decision for one institution AND a benchmark every similar institution can reuse. That double-yield is the design, not a side effect.
Three ways to back the build
Each route is scoped, named, and produces artifacts you can point to. We co-design the scope with every funder.
Fund a sprint or fellowship cohort
A named cohort of fellows completes an applied sprint and ships public deliverables — with your organization credited on the registry and in the artifacts.
Underwrite an evaluation track
Fund a benchmark suite, model-card series, or evaluation report for a sector you care about — published openly, reusable by every institution that follows.
Anchor the applied AI lab
Multi-year backing for the policy-to-production lab: readiness sprints, evaluation infrastructure, governance patterns, and the talent pipeline that runs it.