Causal Inference Decision Framework Director

The mandate

This project will create a governed causal-inference framework for recurring executive decisions, without disclosing the confidential interventions or outcomes in advance. The Director must specify how questions become estimands, how identification assumptions are tested and how uncertainty reaches decisions. The engagement does not operate experiments, build a production platform or make the underlying management choices.

Named artifacts are a decision-question catalogue, estimand template, identification-strategy matrix, assumption and falsification protocol, data suitability assessment, method-selection guide, sensitivity standard, reporting format, three worked pilots and an owner handbook. Each pilot must preserve the difference between prediction, association and causal effect.

Documentation must state the intervention version and assignment mechanism actually observed; a policy label alone is insufficient when implementation intensity or eligibility changed over time.

Four dated milestones govern six months: scope and question selection by 30 October 2026; framework and data assessment by 11 December; three pilots with falsification results by 12 February 2027; final internal reproduction and acceptance by 26 March. A Quantitative Methods Committee is the acceptor after data governance and decision-owner reviewers record challenge.

Final acceptance requires internal analysts to restate one new question as a precise estimand, choose a defensible design, surface seeded positivity and unmeasured-confounding concerns, run appropriate sensitivity analysis and communicate what cannot be concluded. Critical assumption or lineage defects block sign-off.

Management provides approved questions, intervention records, outcome definitions, relevant covariates, policy timing and timely decisions. Consulting owns method, pilots and transfer; management owns data assertions and decisions. Platform construction, new data procurement, operational experimentation, legal evaluation and recurring analysis are excluded.

What you will own

• Translate decision questions into population, treatment, comparator, outcome, horizon and estimand.
• Define method selection across randomised, quasi-experimental and observational designs based on identification conditions.
• Require causal diagrams or equivalent assumption records that expose confounding, mediation, selection and interference.
• Establish balance, overlap, placebo, pre-trend, negative-control and sensitivity checks appropriate to each design.
• Prevent predictive accuracy from being presented as evidence of intervention effect.
• Produce decision reports that separate identified result, assumptions, range, transportability and residual uncertainty.
• Pilot three materially different questions and transfer the approach through an unseen internal case.
• Apply change control to data procurement, platforms, operations, legal work and repeated production.

Candidate qualifications

• Demonstrate leadership applying causal inference to executive decisions rather than academic illustration alone.
• Describe a plausible analysis you rejected because positivity, interference or unmeasured confounding defeated identification.
• Show how a falsification or placebo test materially changed the claimed effect.
• Evidence clear distinction among prediction, association and causal estimand in senior communication.
• Provide a sensitivity analysis that made an uncertain result useful without overstating it.
• Explain how you tested internal analysts on a case not seen during training.
• Show commercial discipline when new data or platform requests expanded project scope.

Working terms and boundaries

• The fee covers six months, three days weekly, four milestones and three worked pilots.
• The Quantitative Methods Committee accepts outputs after an unseen internal reproduction and critical-defect closure.
• Management owns source assertions, intervention choices and decisions; consulting owns named framework artifacts.
• Missing data or decisions enter documented dependency control before dates change.
• Platform build, data purchase, operational experiments, legal analysis and recurring production are excluded.

Application

Applications for this mandate are received in one way only: through the India Board Terminal's application process. It is automated end to end. Your Executive Passport travels to the mandate holder in its confidential form, your answers to the three questions below are read before anything else in your file, and every stage that follows is recorded on your applications page.

There is no address to write to and no intermediary to call. The mandate holder reads what the Terminal delivers and nothing else, which is what keeps the process the same for every applicant and keeps your name out of it until you release it. Applications close on 8 October 2026. Mandate reference QNT-CON-2026-WAW-20.