Building a Data Agent — Part 2: Giving the Model a Workspace
What happened when I gave a data agent one tool, added some back, and compared the results across models.
Read article →I'm a Principal Data Scientist leading data science across BCA and cinch. My interests lie in agentic AI, data science, machine-learning platforms, and decision systems.
This is where I write about building reliable AI, technical leadership, and the ideas that shape how I work.
What happened when I gave a data agent one tool, added some back, and compared the results across models.
Read article →I tried planners, reviewers and more specialised tools. The results made me question how much structure was helping the agent answer data questions.
Read article →An honest comparison across weekly prediction, smaller training sets, broken features, latency, API usage, and the awkward gap between benchmark accuracy and production readiness.
Read article →