How Noora Health made LLM emergency triage auditable on a WhatsApp service handling 50,000 medical queries a month: the model extracts symptoms, and clinician-written rules decide. Recall rose from 0.57 to 0.81.
Aman Dalmia and others
The 4-level AI evaluation framework applied to real deployments, presented with APHRC at the Deep Learning Indaba in Lagos.
Watch Evaluating and Improving AI Agents with Calibrate A step-by-step walkthrough of Calibrate using a form-filling voice agent: structured model tests, trading off latency, cost and quality, aligning LLM judges with human experts, monitoring agents in production, and evaluating speech-to-text and text-to-speech end to end.
Watch Turning Claude, Codex and Cursor Into an AI Eval Agent How to wire Calibrate into Claude, Codex and Cursor so your coding agent can find where an AI product is failing and drive the fixes.
Watch Evaluating AI for Social Impact A practitioner's workshop on the 4-level AI evaluation framework, taking model evaluation and A/B testing from first principles to a live demo. Recorded at the Africa Evidence Summit in Addis Ababa.
Watch AI Evals Office Hours Open office hours where non-profits bring the practical AI evaluation problems they are actually hitting, and we work through them live.
A working record of the behavioral science behind TAF's portfolio: 16 studies across 9 partner orgs, publishing the nulls alongside the wins.
Edmund Korley and others
Watch Learn Enough Codex To Be Dangerous A hands-on workshop on agentic AI with Codex, run through the Agentic Workflows Club.
3 files · ZIP Download GenAI for Data Analytics When to use, and when not to use, GenAI for data analytics, and how to keep analysis safe and collaborative. Workshop deck, instructions, and a sample dataset.
Watch AI Agent Evaluation Tutorial A step-by-step walkthrough of how to evaluate AI agents, from Aman's Calibrate tutorial.
Watch Learn Enough Claude To Be Dangerous A hands-on workshop on using Claude Code and Cowork for agentic AI workflows.
A master slide deck for adapting the AI evaluation playbook into workshops, trainings, and talks.
Watch AI 101 
Watch Model Eval for Leaders An overview of AI model evaluation within the 4-level framework, for leaders who care about real developmental outcomes, not just engagement.
A quick introduction to the 4-level AI evaluation framework, used when onboarding new accelerator orgs.
How leaders should think about data and metrics, used when onboarding new accelerator orgs.
A high-level walkthrough of the AI benchmarking process for health AI grantee organisations.
Zezhen Wu and others
3 files · ZIP Download AI4GD 2025 Model Eval Workshop A zipped set of the notebook, working sheet, and knowledge base from the Nairobi AI4GD model evaluation workshop.
James Walsh and others
Temina Madon, Robert On, James Walsh, Zezhen Wu, and others
Robert On and others
Zezhen Wu and others
Zezhen Wu and others
Robert On and others
Richard Sedlmayr and others
Nothing in this category yet.