Senior Data Scientist

We're hiring a Senior Data Scientist to lead end-to-end machine learning model development for [core product area: growth, monetization, trust & safety, recommendations, etc.]. You'll own the entire lifecycle: identifying opportunities, building models, shipping to prod, and measuring impact. This is a hands-on IC role with high autonomy and accountability.

Key Responsibilities

1. Discovery & Scoping

• Work with Product, Eng, and Analytics to uncover high-leverage ML opportunities
• Define problem statements, success metrics, and evaluation strategy upfront
• Perform exploratory analysis to assess feasibility and estimate ROI

2. Model Development

• Build features from structured + unstructured data at scale: logs, events, text, images, time-series
• Select and implement the right approach: regression, classification, clustering, deep learning, LLMs, causal models, etc.
• Run rigorous offline experiments: cross-validation, hyperparameter tuning, error analysis

3. Deployment & Experimentation

• Partner with MLEs/DE to get models into production: real-time APIs, batch pipelines, edge
• Design A/B tests and interpret results. Make ship/no-ship calls based on data
• Build guardrails: latency, fairness, reliability, and cost requirements

4. Production & Iteration

• Implement monitoring for feature drift, prediction drift, and performance decay
• Own model maintenance: retraining, tuning, deprecation
• Close the loop: use prod learnings to inform v2, v3 of the model

5. Org Impact

• Raise the bar: code reviews, tech talks, reusable tools, documentation
• Mentor mid-level DS. Be the person others come to for ML design questions
• Evangelize data-driven decision making across the company

Qualifications

You Must Have:

• 5+ years building ML models end-to-end with proven business impact. You've shipped, not just prototyped
• Fluent in Python and SQL. Strong grasp of numpy, pandas, scikit-learn. Experience with PyTorch or TensorFlow
• Solid ML theory: can explain regularization, boosting, embeddings, and transformer basics without notes
• Experience with large datasets: Spark, Presto, BigQuery, or similar. You know when to sample and when not to
• Track record of designing experiments and measuring incremental lift, not just accuracy
• Product mindset: you care about users and business metrics as much as model metrics