Senior Data and AI Scientist

Role overview

Responsibilities

• Develop models for recommendations, predictions, classification, and other product features; explore data and engineer features.
• Evaluate performance, run experiments, and prototype algorithms in NLP, representation learning, and time-series forecasting.
• Generate insights through analytics, segmentation, and statistical analysis and communicate findings to teams.
• Contribute to A/B testing, model validation, dashboards, and monitoring for model health and business KPIs.
• Deploy models with product and AI engineering teams and integrate them with backend services and APIs.
• Create reproducible training pipelines and notebooks; contribute reusable model components, feature stores, documentation, and internal tools.

Candidate qualifications

• 4–6 years in data science, ML engineering, or applied AI; SaaS or technology product experience is ideal.
• Experience building customer-facing or business-impacting models and familiarity with Git, code reviews, and ML CI/CD.
• Bachelor’s or master’s degree in Data Science, Computer Science, Statistics, Engineering, or a related field.
• Strong communication skills, ability to work independently and collaboratively, and awareness of fairness, bias, and explainability in ML.
• Technical tools include Python, SQL, Jupyter, Git, ML frameworks and data tools; experience with one of GCP, AWS, or Azure ML is required.
• Experience with robustness, interpretability, inclusivity, production monitoring, advanced ML areas, Docker/Kubernetes, feature stores, or BI tools is a plus.

Technical skills

• ML frameworks named include scikit-learn, XGBoost, LightGBM, and TensorFlow or PyTorch at a basic level.
• Data tools named include Pandas, NumPy, DBT, BigQuery, and Snowflake or similar.
• MLOps tools include MLflow, Airflow, Kubeflow, and optional feature-store tools such as Tecton or Feast.
• Cloud examples include GCP Vertex AI, AWS SageMaker, and Azure ML; Docker and Kubernetes are a plus.
• Visualization examples include Seaborn, Matplotlib, Plotly, and optional BI tools such as Looker, Power BI, and Tableau.
• Model-evaluation knowledge includes overfitting, bias, variance, cross-validation, and A/B testing.

Work arrangement

• Benevity describes its approach as flexible hybrid, with no fixed in-office attendance requirement for people near its offices.