Senior Applied / Decision Scientist who names the decision criterion before running any causal analysis on data nobody randomized. Yannis covers causal inference on observational data (propensity score matching, difference-in-differences, instrumental variables, regression discontinuity), turning a vague "should we do X" question into a measurable analytical question with a clear decision criterion, and optimization or simulation for resource allocation decisions. Who it's for Analytics and strategy teams who need a defensible causal estimate from historical data, plus a stress-tested allocation plan, not a naive before/after comparison dressed up as proof. Key capabilities Fuzzy business asks translated into a measurable question with a stated outcome and decision criterion Identification strategy (PSM, diff-in-diff, IV, RDD) chosen from how the treatment group actually came to be treated Identifying assumptions tested explicitly: covariate balance, parallel pre-trends, instrument relevance, cutoff manipulation Effect estimates delivered with an honest confidence range and a named list of remaining confounders Resource-allocation optimization or simulation stress-tested against the uncertainty in the underlying estimate How to use it Paste Yannis's SKILL.md into your Claude Project Instructions (or any AI system prompt), then describe your Applied & Decision Scientist problem. Works with Claude, ChatGPT, and any AI chat. Under 2 minutes to install.
