ImpactAs a Senior Data Scientist on the Search Science team at Shipt, you will responsible for the algorithms that determine what customers see every day. You will be working to re-architecting Shipt's search stack to be LLM-native and value-aware.
In this role you will build a state-of-the-art retrieval and ranking system from the ground up. You will partner with Senior Scientists and ML Engineers to push the boundaries of what a search engine can do. Search is the front door of Shipt-improvements you make here are immediately visible to leadership and impactful to the bottom line.
How you can make an impact:
Unified Ranking: Help design and implement multi-task learning (MTL) and multi-objective models that don't just optimize for clicks, but solve for complex trade-offs between organic relevance, ad revenue, and operational efficiency.
In Depth Search Analysis: Identify opportunities, explore datasets, and report out findings to senior level leaders in the company which will influence our Search roadmaps.
Deep Personalization: Move beyond generic relevance by integrating rich user signals through things like purchase history, dietary constraints, and real-time intent. Intentionally optimize our search system to predict not just what is relevant, but what is relevant for each customer.
LLM-Native Search: Move beyond keyword matching by integrating Large Language Models into the core search loop. Work on projects such as generative retrieval, query expansion, and semantic understanding to handle complex, intent-driven user queries.
Hybrid Retrieval Systems: Develop and tune sparse (lexical) and dense (vector) retrieval layers to maximize recall across our massive catalog, ensuring we surface the right products even for long-tail queries. Explore generative retrieval methodologies and their utilization.
What You'll Need to Be Successful- Bachelor's degree in a STEM field (Computer Science, Statistics, etc.) with a focus on machine learning
- 3-5 years of experience applying Machine Learning to real-world problems
- Strong coding fluency in Python and SQL
- Deep familiarity with Deep Learning frameworks (PyTorch or TensorFlow) and their application to information retrieval, ranking and/or NLP
- A solid grasp of Learning-to-Rank (LTR) concepts including rich feature engineering
- Previous experience with Search or Recommender Systems in e-commerce settings
- Experience with Multi-Objective Optimization: Understanding of the challenges of optimizing for multiple conflicting metrics (e.g., Relevance vs. Revenue)
Pay Range: $115,000 - $170,000Please note that the salary range above is a guideline, and individual total compensation will vary based on factors such as qualifications, skill level, competencies, and work location.Skills & EducationThis list includes key skills used in this job but is not inclusive of all skills needed for the role. Please see any required education below.
Data Science, Machine Learning (ML), Search EnginesBachelor's Degree or equivalent experience | Required
Work ArrangementTo foster connection while offering continued flexibility, hybrid team members have the following in-office expectations:
Hybrid roles in Birmingham, AL typically work in-office at least 2 days per week, with core in-office days on Wednesdays and Thursdays.
Hybrid roles in Minneapolis, MN typically work in-office at least 1 day per week on either Tuesday, Wednesday, or Thursday.
Hybrid roles in San Francisco, CA typically work in-office at least 1 day per week between Monday and Thursday.
Certain roles may require in-office presence on a full-time basis. Please work with your recruiter to learn more about the classification of this role.
Employees (and eligible family members) are covered by medical, dental, vision and more. Employees may enroll in our company's 401k plan. Employees will also be eligible to receive discretionary vacation for exempt team members, paid holidays throughout the calendar year and paid sick leave. Other compensation includes eligibility for an annual bonus and the potential for restricted stock units based on role.
