Discover's highly-scalable serving system needs to design, build, and ship production machine learning models and exploration algorithms for retrieval, scoring, and ranking to improve content helpfulness for users.
Requirements
- 5 years of experience with software development in C++, production systems, and machine learning.
- 1 year of experience with software design and architecture.
- 3 years of experience with full stack development, across back-end such as Java, Python, GO, or C++ codebases, and front-end experience including JavaScript or TypeScript, HTML, CSS or equivalent.
- Experience implementing and shipping deep learning models (e.g., two-tower or other types of deep neural networks) for retrieval or ranking in a major recommendation product.
- Experience developing accessible technologies and with algorithms designed to improve content and solve the exploration/exploitation tradeoff.
- Experience with the infrastructure that powers modern search and recommendation systems
Responsibilities
- Design, build, and ship production machine learning models and exploration algorithms (e.g., contextual bandits) for retrieval, scoring, and ranking in Discover's highly-scalable serving system.
- Work across the full technical stack in C++ and Python.
- Prototyping and implementing machine-learning models
- Building real-time data pipelines
- Integrating with core Search infrastructure
- Leveraging serving systems
- Running experiments at scale.
Other
- Participate in, or lead design reviews with peers and stakeholders to decide amongst available technologies.
- Direct the full project lifecycle by analyzing results, iterating based on metrics and product feedback, and applying deep user-centric judgment to define what makes content truly helpful for users.
- Thrive in ambiguity by navigating speculative problem spaces and translating high-level product goals into concrete technical solutions.
- Bachelor’s degree or equivalent practical experience.
- 3 years of experience testing, maintaining, or launching software products.
