TThamu Mnyulwa·Jul 22MLflow Autologging and Zero-Config TrackingHow to get useful experiment tracking with one line of instrumentation, and where to add explicit control.6 min read·thamu.hashnode.dev
TThamu Mnyulwa·Jul 19Reproducible Experiment Tracking with MLflow and UVA practical workflow for pairing MLflow runs and model versions with a locked Python environment.7 min read·thamu.hashnode.dev
TThamu Mnyulwa·Jul 14Productionising Generative AI with MLflow 3.x: Tracing, Evaluation, and Prompt OptimisationA practical blueprint for taking RAG and agentic systems from demos to production.12 min read·thamu.hashnode.dev1A
TThamu Mnyulwa·Jul 14GenAI Evaluation with mlflow.genai.evaluate(): Beyond AccuracyA practical evaluation harness for RAG systems, LLM judges, custom scorers, and prompt version comparisons.10 min read·thamu.hashnode.dev
VViktor Berthelius·May 19Evaluation Ownership: who answers when the model improves output but worsens the decisionEvaluation breaks when everyone tracks output quality and nobody owns the business consequence of a worse decision.3 min read·brthls.com
TThamu Mnyulwa·Apr 30Sentinel Health: Winning the Google Build with AI Buildathon Southern Africa 2026A spec-driven 60-minute build from a rural-health brief to a backend, frontend, and visible cloud MVP.8 min read·thamu.hashnode.dev
VViktor Berthelius·Apr 17Executive Review Stack for AI: the weekly CEO stack for real governanceAI governance becomes real when the CEO reviews decision quality, escalation load, and value leakage every week.3 min read·brthls.com
VViktor Berthelius·Apr 7Agentic Category Design: structuring pages and claims for agent shortlistsAgents shortlist products through category clarity, evidence structure, and claim legibility before any human conversation happens.2 min read·brthls.com
VViktor Berthelius·Apr 3Model Routing as Governance: the policy layer behind model choiceModel routing becomes governance the moment teams stop choosing models by taste and start routing by policy.2 min read·brthls.com
VViktor Berthelius·Mar 19Agent Orchestration 2026: LangGraph, CrewAI and the False Sense of ScaleOrchestration without decision governance is automation theater: more tools, less control.2 min read·brthls.com