Clinical data and ADaM
Take a messy supplied SDTM extract and produce an ADaM dataset with a define.xml that validates, and traceability you can demonstrate.
You finish with: An ADaM dataset, a passing define.xml, and a traceability note.
A paid training programme for statisticians, programmers and engineers who want to work on clinical data but have never seen what a submission actually demands.
We put that at the top rather than in the small print. You pay a fee, you get eight weeks of structured work on realistic clinical problems, review from engineers who do this for clients, and a certificate. You do not get a salary, a contract or a guaranteed interview. What you get is a portfolio artefact and the vocabulary our open roles ask for.
Take a messy supplied SDTM extract and produce an ADaM dataset with a define.xml that validates, and traceability you can demonstrate.
You finish with: An ADaM dataset, a passing define.xml, and a traceability note.
Take a protocol summary and write the estimand and the analysis section of a SAP, including how intercurrent events are handled.
You finish with: A written estimand and SAP section, reviewed line by line.
Qualify a small R environment: pin dependencies, document package qualification, and produce the evidence a validation lead would ask for.
You finish with: A reproducible environment with its qualification evidence.
Design and run a small benchmark against an agent, then write up what it actually shows and what it does not.
You finish with: A benchmark harness and an honest write up of the results.
Built for people with a degree to finish or a job to keep. Two weeks grounding in the domain and the tooling, four weeks building the track project in reviewed weekly increments, one week defending your assumptions, one week writing it up cleanly.
Places are capped at twelve per cohort because review time is the real constraint.