Staff Data Scientist, Marketing Science ($200k - $300k salary + bonus)

Baton Corporation Ltd•New York, NY

3d•$200,000 - $300,000•Onsite

About The Position

Baton Corporation is hiring a founding Staff Data Scientist in Marketing Science to build and own the scientific methods behind paid growth. The role will determine how the company allocates its multi-million-dollar, always-on acquisition budget across channels, using causal measurement, response curves, customer-value models, and optimization. There is no inherited methodology; the hire will design the approach from first principles and continuously recommend where additional budget is likely to generate the most incremental profit, while communicating the confidence behind those recommendations. This is a hands-on builder role. The scientist will write models, design experiments, and work closely with the people managing acquisition budgets. Data Engineering owns data pipelines, while this role owns the science. The position works with the Head of Paid Acquisition & Performance Marketing and partners with Data Engineering. Baton describes its environment as in-person, intense, fast-paced, resource-lean, and characterized by long and sometimes unusual hours, minimal bureaucracy, daily production releases, and end-to-end individual ownership. New York is preferred; exceptional candidates may be considered for remote work with strong US Eastern-hours overlap and regular travel to New York. The position has a base salary range of $200,000–$300,000, depending on experience and scope, plus a generous performance-related bonus. Equity or token participation may be offered depending on level. Benefits and relocation support are to be confirmed for this role.

Requirements

Advanced quantitative training: a PhD, MSc, or equivalent depth in mathematics, statistics, econometrics, physics, or operations research.
At least 4 years of hands-on experience, with PhD research counting toward experience.
Track record of building marketing mix or advertising response models that influenced real budget decisions, with results the candidate can explain and support with evidence.
Hands-on causal inference experience, including designing, running, and analyzing incrementality experiments and addressing discrepancies between experimental results and model outputs.
Deep Bayesian modeling expertise, including hierarchical models, justified priors, identification, posterior diagnostics, and computational methods for fitting models efficiently.
Strong Python or R skills for production, including writing model code directly.
Commercial judgment focused on incremental profit, LTV, CAC, and payback, and ability to turn model results into actionable recommendations for non-specialists.
Scrappy, high-energy operator mindset and ability to work in a fast-paced, resource-lean environment while building processes and methods.

Nice To Haves

Contributions to open-source measurement tools such as PyMC-Marketing, Meridian, Robyn, LightweightMMM, CausalPy, Stan, or GeoLift.
Mobile app user acquisition experience, including SKAdNetwork, AppsFlyer or Adjust, and Apple Search Ads.
Crypto experience as a retail trader or through work at a crypto startup or larger crypto company.

Responsibilities

Build a Bayesian marketing mix model incorporating adstock, saturation, channel-level response curves, and uncertainty, using a defensible modeling framework.
Design and analyze geo-experiments, conversion-lift tests, holdouts, and synthetic-control studies to measure incrementality and calibrate models.
Estimate marginal ROAS and marginal CAC by channel, perform constrained budget optimization, and provide risk-aware budget recommendations for each planning cycle.
Build LTV, retention, and payback models tied to user activation and trading activity to optimize for profit rather than platform-reported ROAS.
Develop defensible measurement approaches for newer or opaque channels where lift tooling is limited or unavailable.
Work with the Head of Paid Acquisition & Performance Marketing to establish test roadmaps, define success metrics, and determine what to scale or stop.
Ship versioned, tested, reproducible, production-quality code and partner with Data Engineering on required model inputs.
Communicate uncertainty in clear decision memos that founders and operators can act on, including when further testing is needed.
Audit current measurement, launch a baseline response model, and design initial incrementality tests.
Develop an experiment-calibrated model that produces channel response curves, marginal returns, and budget recommendations with confidence ranges, and use it in the budget cycle.
Drive measurable incremental profit through model-informed budget reallocations and establish an always-on test-and-learn program across existing and new channels.