We are a bold marketing agency built for brands who need results now. Our data-fueled strategies help brands rise above the competition with culturally relevant high-performance creative solutions that generate energy, action and revenue. We are seeking an experienced Data Scientist/Statistician to work in our Pittsburgh, Chicago, or Boston office.
This is a great opportunity for the Data Scientist/Statistician to use modern statistical techniques (causal modeling, hierarchical/multi-level regression, Bayesian statistics, state-space time series, MCMC, HMC) to help our clients answer their most pressing business challenges and questions.
This position reports to the VP, Director of Data Science.
- Deep understanding of parametric statistics and causal statistical modeling
- In-depth knowledge of advanced modeling techniques like…
- Causal modeling, spurious correlation, confounding variables, collider bias, post-treatment effect bias, directed acyclic graphs (DAGs)
- Hierarchical/Multi-level Regression (random, fixed, mixed effects)
- Econometrics, state-space models, decompositions, forecasting, log-linear models, elasticity
- Estimation techniques MCMC (Metropolis-Hastings, Gibbs sampling, Hamiltonian MC)
- Experimental design, measurement error, etc.
- Foundational Data Science and data understanding skills
- Querying data bases, connecting to APIs, finding open-source data
- Restructuring, imputing, feature engineering, “cleaning” (we all “clean” our own data and believe it’s a fundamental part of understanding the data used in the analyses)
- Robust exploratory data analysis practices (visualization of marginal and joint distributions as well as time series)
- Ability to work with both small data sets and large data sets alike
- Python or R or Julia skills are a must
- Secondary analytical skillsets (nice-to-haves)
- Unsupervised Machine Learning (clustering, dimension reduction/compression techniques like PCA, KPCA, T-SNE & UMAP, etc.)
- Semi-supervised learning
- Classification methods and scoring functions (SVM, RF, NN, etc.), especially unbalanced classes
- NLP (topic modeling specifically)
- Reinforcement Learning (think pricing optimization)
- Cloud environments (we use AWS, but experience with Azure or Google Cloud is cool with us)
- GitHub experience
- Soft Skills
- Capacity to identify and understand business questions and hypothesize appropriate data science solutions
- Ability to participate in meetings with clients, but no need to lead them
- Communication skills to ask questions of clients and internal partners to clarify data definitions, data structure and data generation biases
- Excitement to apply advanced methods to business and marketing scenarios
Some of the questions you’ll be answering:
- What was the impact of our pricing changes on overall sales and profitability?
- What would happen if we shift our media spend channels?
- What is the optimal level of advertising spend by month for the next 6 months?
- What is the optimal pricing for each product in a set of SKUs?
- Creative, gritty, scrappy, ability to flourish under ambiguity, and passionate to learn
- Fits well into results-oriented culture with little-to-no red tape
- Ability to explore multiple challenges and models concurrently
- Passionate about Math, Statistics, Data Science, and Machine learning
- 4+ years in Statistician, Data Scientist, Machine Learning, role or similar
- Role focused on business, strategy, marketing, media
How this position is different:
- No red tape around techniques or solutions. I.e. There is extensive creative freedom for building and solving the problems at hand without unnecessary constraints on modeling techniques—as long as those modeling techniques answer the question
- Fast-paced with exposure to clients from a variety of industries (retail, quick serve restaurants [QSR], food and alcohol, tourism, education, medical device, medical services, etc.)
- Culture that fosters continual learning, curiosity, innovation
- Work schedule and location flexibility
- Travel to clients or other 9Rooftops offices when appropriat