About
Cody Bakken
I lead data, analytics, and AI as an executive who still ships. I've built these functions from zero twice, and I stay hands-on in the code while setting the strategy.
I'm VP of Data & Analytics at Derivita, where I serve on the leadership team and report to the CEO. I own analytics, data, and AI across both the customer-facing product and the internal business — a function I built from the ground up and run on a deliberately lean, AI-augmented model: a small team paired with AI coding agents, shipping production analytics end-to-end at a fraction of typical headcount and cost.
That's produced assessment-level predictive analytics that contributed to major multi-year renewals, a company-wide internal analytics platform that became the single source of truth, the RAG retrieval and LLM-judge system behind automated standards tagging, and the analysis behind an independent efficacy study of 1,647 students. Before Derivita, as VP of Data Science at InMoment, I founded and led an 11-person data science organization across three countries, shipping ML and NLP products that added millions in recurring revenue.
How I think about the work
- Measurement is the whole game. An output that looks right isn't the same as one you can defend and measure.
- Good AI rests on well-curated data. Weak data foundations are where most AI projects break.
- AI is a tool. The fundamentals of good data work still decide whether it delivers.
- An executive who still ships. Close enough to the code to know what's real.
I write here about agentic analytics: getting LLM agents to do real analytical work, and the harder problem underneath it — knowing when the answer is wrong.