Staff Data Scientist

Toast Toast · Enterprise · United States · Remote · R & D : Engineering : Commerce

Staff Data Scientist at Toast to lead the design and development of scalable ML systems for restaurant business use cases like recommendation, forecasting, and personalization. Role involves full ML lifecycle ownership, from problem framing to deployment and monitoring, with a focus on influencing roadmap, setting best practices, and mentoring junior scientists. Requires strong Python, SQL, ML frameworks, cloud platform experience, and software engineering principles.

What you'd actually do

  1. Own the full machine learning lifecycle—from problem framing and data exploration to modeling, deployment, and monitoring—for mission-critical initiatives.
  2. Design and implement advanced ML and statistical models that improve product performance, operational efficiency, or customer insights.
  3. Collaborate with engineers, product managers, and business stakeholders to define project scope, success metrics, and integration strategy.
  4. Guide architectural decisions, set modeling standards, and champion best practices for experimentation, validation, and productionization.
  5. Mentor other data scientists and raise the technical bar through design reviews, feedback, and sharing domain expertise.

Skills

Required

  • Python
  • SQL
  • scikit-learn
  • PyTorch
  • TensorFlow
  • AWS
  • SageMaker
  • Athena
  • Glue
  • DynamoDB
  • Bedrock
  • statistical modeling
  • machine learning
  • model evaluation
  • distributed data processing
  • training
  • real-time inference
  • ML Ops frameworks
  • experimentation
  • A/B testing
  • causal inference
  • real-time decision systems
  • software engineering principles
  • modular design
  • version control
  • testing
  • CI/CD

Nice to have

  • advanced degree in Computer Science, Statistics, or a related STEM field
  • MLOps tooling for monitoring, drift detection, retraining, and explainability
  • fine-tuning LLMs
  • reinforcement learning from human feedback (RLHF)

What the JD emphasized

  • proven track record of delivering production ML systems that drive measurable impact
  • translating ambiguous problems into well-scoped ML solutions
  • real-time inference
  • ML Ops frameworks
  • cloud platforms

Other signals

  • deliver production ML systems
  • translate ambiguous problems into ML solutions
  • ML Ops frameworks
  • cloud platforms