Data Scientist
About The Position
At Kissterra, we are revolutionizing the insurance industry with the world's first Marketing OS - a comprehensive solution tailored to the unique marketing and distribution needs of insurance organizations.
Our intelligent solutions combine the power of AI, machine learning, and real-time data to reshape marketing strategies and drive profitable growth for top insurers.
Founded in 2015 and headquartered in Israel, Kissterra partners with leading US insurance companies, helping them uncover hidden growth opportunities and optimize their marketing budgets with unparalleled precision and efficiency.
You'll be Kissterra's first data scientist. You'll own the models that value prospects and drive our automated bidding, and you'll shape how data science works here: which problems to take on, and how to ship and measure the work. There's no team lead through influence, working across Product, Data Engineering, Platform Engineering and the business teams.
What You'll Do:
- Value prospects: predict conversion and bind probability, and expected value per lead, across products and channels.
- Bid and price: design and tune bidding strategies that balance margin
- Test before shipping: backtest changes on historical data, then confirm them with live A/B tests.
- Ship to production: work with engineering to put models into our scoring services, then monitor accuracy and financial impact.
- Set up data science: agree a roadmap with Product and R&D, define success metrics, and set modeling standards. Work with Data Engineering on model-ready datasets.
Requirements
- 5+ years building ML or pricing models and owning them end to end in production.
- Has done similar work with little existing structure. Experience is a strong plus.
- Strong Python (pandas, NumPy, scikit-learn) and gradient-boosted models (XGBoost, LightGBM or CatBoost), including probability calibration.
- Strong SQL on large transactional datasets.
- Familiarity with Git and source control workflows
- Solid grounding in probability, statistics, experimentation and causal
- Works daily with AI coding agents (e.g. Claude Code, Cursor) and critically reviews their output.
- Explains model results and trade-offs clearly to non-technical stakeholders.
- B.Sc. or higher in a quantitative field, or equivalent experience.
- Ad tech / RTB or dynamic pricing preferred- Advantage
Nice to Have:
- Experience with insurance lead generation or performance-marketing platforms.
- ClickHouse, Airflow, Apache Iceberg.
- Deploying models as containerized services (Docker / Kubernetes).
- Hands-on with AWS S3, Athena, EMR, Spark (PySpark) and dbt