Research and writing on data, AI, migrations, and governance.

EMR Migration With Zero Data Loss

Toronto specialty hospital migrates legacy EMR to Epic

How a 3,000-patient specialty hospital moved off its legacy EMR onto Epic with no material data elements lost

Your Hospital's Next Board Report Depends on the System Behind It

From fragile pipelines to recoverable analytics

How Momo Analytics consolidated clinical and operational data into a governed warehouse with automated pipelines, Metabase reporting, monitoring, and disaster recovery

A Hospital Analytics Platform Built to Survive a System Failure

Toronto hospital evolves from spreadsheets to a governed data platform

How hospital leadership and Momo Analytics turned a data strategy into a governed platform with automated pipelines and a tested one-hour recovery process

A COVID Decision Platform Used Across 40+ Countries

Bayesian outbreak tracking for a United Nations agency

How a Bayesian reproduction-number model, built for a UN agency's pandemic response, gave national and regional decision-makers a daily read on outbreak trajectory instead of a lagging case count.

The Vibe Coding Ceiling

Vibe coding is a real productivity gain for prototypes and throwaway scripts. The data on production codebases tells a different story- one where defects compound, the productivity deminishes and the fix is rather obvious.

The Cloud You Might Not Need

Most companies are running workloads on cloud infrastructure sized for a scale they'll never hit. The data on data size, cost, and cost recovery says a single well-run server does the job for far longer than the default architecture we choose nowadays.

Finance and AI- where's the value?

The industry conversation about AI in finance centers on portfolio judgment and alpha. However, research points toward engineering work with clear feedback loops, and away from allowing LLM to make judgement calls under uncertainty.

Pitfalls to watchout for in an EMR Migration

A brief account of how health records can be lost, and how to catch it before go-live

Most EMR migration failures aren't in the data model - they're in the last 5% of records that don't match the scope. Notes from a legacy-to-Epic migration with zero material data loss.