Текст заказа
What Makes Us a Great Place to Work
We are proud to be consistently recognized as one of the world’s best places to work. We are currently the top ranked consulting firm on Glassdoor’s Best Places to Work list and have earned the #1 overall spot a record seven times. Extraordinary teams are at the heart of our business strategy, but these don’t happen by chance. They require intentional focus on bringing together a broad set of backgrounds, cultures, experiences, perspectives, and skills in a supportive and inclusive work environment. We hire people with exceptional talent and create an environment in which every individual can thrive professionally and personally.
Who You’ll Work With
Growth AI Performance System is Bain’s integrated suite of agentic decisioning tools giving large organizations always-on intelligence across their most important growth decisions. It sits at a rare intersection few organizations can genuinely offer: the analytical depth and client access of a top-tier consultancy, the technical ambition of a product company, and the commercial stakes of both.
You will lead the statistical and modelling practice behind the platform -- the software that helps Fortune 100 to 500 companies answer where to invest, where to focus, and how to win across marketing ROI, brand strategy, and pricing and commercial analytics. You will work inside Bain’s persistent product team, alongside Tech Leads, Full-stack Engineers, and AI Engineers, with direct feedback from case teams and client engagements flowing back into the models.
The platform has two layers working together: the Agentic Core and three product modules, each solving a distinct commercial problem. The build lifecycle is uniquely connected to the real world. As case teams work alongside clients across industries and geographies, that feedback flows directly back to engineering and gets woven into the product. Every client engagement makes the suite smarter, broader, and more resilient.
This is a startup inside an established company. You get genuine early-stage ownership -- real influence over architecture and product direction -- backed by Bain’s global scale, client relationships, and leadership commitment. The pace of a startup with the runway of a global firm.
We are looking for a research-calibre statistician who goes past the first defensible answer when the data and the decision require it -- someone who can derive a model on a whiteboard, implement it correctly in code, validate it rigorously, and defend it in a client room. This is applied statistics in production, at scale, with real users depending on what you build.
Where You’ll Fit Within the Team
The Head of Data Science owns the statistical and analytical standard for every Growth AI Performance System product module. You sit above module-level Data Scientists, working across blended product and case teams. You are the final word when a modelling question exceeds the standard playbook -- and you are expected to personally lead the hardest, least-precedented problems, not only set direction for others.
You will work across the platform and product modules focused on marketing ROI, brand or pricing strategy. The problems span marketing mix modelling, brand equity measurement, price elasticity, promotional lift, and ranging and assortment -- hard analytical domains with decades of methodology behind them and enormous client-specific noise in front of them. You must know that literature well enough to depart from it when it does not serve the client, and produce outputs credible in both directions: to a methodologically literate peer and to a CMO or Category Director who needs to act.
Analytical outputs here are not research artefacts. They are the core of what the product delivers and what clients trust. The standard you hold the team to is the standard the whole product is measured against.
This role requires a combination that is genuinely rare: research-grade statistical depth, the software discipline to implement models that engineers can maintain, and the client communication skills to present methodology, defend assumptions under live challenge, and adjust the model narrative when the room requires it.
What You’ll Do
_Analytical leadership and frontier modelling_
* Set and own the statistical standard across all Growth AI Performance System modules: define what methodologically sound means from problem framing through production handoff, and hold every module’s work to that bar
* Personally lead the hardest modelling problems -- where standard Bayesian, time-series, hierarchical, or domain-specific playbooks do not cleanly apply -- including cases requiring custom likelihoods, non-standard priors, partial identification analysis, or cross-model reconciliation
* Take models written as equations -- likelihood, priors, constraints, forward maps -- and implement correct, tested, reproducible estimators in Python using probabilistic programming frameworks; derive estimators from first principles and diagnose identifiability, partial identification, and convergence
* Know when a model cannot support the client question -- not just when library defaults fail silently -- and have the analytical authority to reframe the problem rather than deploy a model that looks fitted but is not identified
* Design and enforce inference and validation standards across all modules: MCMC and VI diagnostics, prior sensitivity, posterior predictive checks, simulation-based calibration, parameter recovery on synthetic ground truth, out-of-sample validation, back-testing, and documented confidence bounds before any output reaches a live client environment
* Own how statistical outputs become structured, versioned data products for engineers and agentic workflows: clear schemas, configuration, interpretable uncertainty measures, and explicit confidence bounds that do not require the original modeller to interpret
* Build and maintain a cross-model reconciliation framework for heterogeneous models that share derived commercial quantities: agreement diagnostics, guarded combination of posterior beliefs, and rules for when reconciliation versus refit is required
* Drive methodological innovation across hierarchical Bayesian methods, causal and quasi-experimental inference, time-series and state-space methods, discrete choice modelling, and regularisation and shrinkage -- bringing improvements into the product when they materially improve client decisions
_Domain modelling across product modules_
* Marketing mix and media ROI: ad stock and saturation structures; incrementality calibration via geo experiments, holdouts, or quasi-experimental designs; attribution under collinearity; budget optimization under uncertainty; integration with platform and clean-room data limitations
* Brand equity and consumer choice: latent brand constructs; measurement error models; survey-to-sales linkage; panel representativeness and bias correction; fusion of tracking, behavioral, and sales data
* Pricing, promotional lift, and assortment: promotional lift decomposition; cannibalization and pass-through; elasticity estimation under endogeneity; ranging and assortment under constraints and cross-SKU substitution
* Stay current with domain data sources and their modelling implications -- media platform data, brand tracking and panel data, point-of-sale data, pricing and promotional data -- including coverage gaps, quality limits, and when upstream data issues are driving misleading outputs
_Team leadership and stakeholder engagement_
* Build and lead the data science practice: hire, mentor, and develop module-level Data Scientists; set the tone for a team that hunts for the best defensible answer rather than the most convenient one
* Present methodology, outputs, and recommendations directly to senior client stakeholders -- translating complex statistical concepts into decision-rele