Senior Data Scientist

Role Overview

Join Smartsheet as a Senior Data Scientist II on the Business Intelligence team. You will build ML models and AI sub-agents to improve growth, monetization, efficiency, and retention across the customer lifecycle.

Responsibilities

• Design and ship AI sub-agents that operate throughout the client lifecycle using:
  • predictive models
  • retrieved context
  • LLM reasoning
• Build predictive and prescriptive models supporting use cases like:
  • churn risk
  • growth and adoption
• Develop the data foundations and knowledge layer used by sub-agents, with a focus on:
  • responsible aggregation
  • privacy-preserving design
• Drive end-to-end delivery: framing problems → building models/sub-agents → production shipping
• Collaborate primarily with Product and Engineering

Requirements

• Deep applied ML expertise across traditional ML and deep learning, including:
  • Gradient boosting
  • Regularized linear models
  • Transformer-based sequence models
  • Foundation model embeddings
  • Causal ML
  • Contextual bandits
  • Offline RL
• Proficient in SQL and Python; comfortable with large-scale ML/LLM tooling such as:
  • Spark / Databricks / Snowflake (or equivalents)
  • ML frameworks: PyTorch, scikit-learn, XGBoost/LightGBM
  • Visualization: Tableau (or similar)
• Strong causal inference for intervention/lifecycle modeling (e.g. uplift modeling, difference-in-differences, propensity scoring, synthetic control)
• Experience modeling SaaS customer lifecycle outcomes: churn, expansion, adoption, plan health, lead/account scoring, and business fluency in SaaS metrics (NRR, GRR, ARR, cohort economics)
• Hands-on experience taking LLM/agent-based systems to production (tool use, retrieval, multi-step reasoning, evaluation, guardrails)
• Strong statistics and experimental design background: hypothesis testing, power analysis, multiple comparisons, sequential testing, quasi-experimental methods
• Experience operating ML in production: feature pipelines, monitoring, drift detection, retraining cadence, and batch vs real-time trade-offs
• Production pragmatism: latency, cost, monitoring, drift, hallucinations, and failure modes for models/sub-agents
• Ability to research and learn new technologies and thrive in both independent and cross-functional environments

Education & Experience

• Bachelor’s degree + 8+ years experience (or 10+ years); advanced degree preferred (Statistics, CS, ML, Economics, Operations Research, or similar)