DataJobs.io
← Back to all jobs

Job Description

Discord’s Engagement ML team is focused on helping people connect faster on the platform. In this role, you’ll work within the Growth organization’s onboarding workstream to build end-to-end recommendation and cold-start personalization systems that improve new-user retention by matching new users with servers, people, and content.

What you’ll do

  • Design, build, and scale recommendation systems that link new users to communities and content, covering the full pipeline from candidate generation and ranking to online serving.
  • Solve cold-start personalization for users with little or no history on Discord, using limited signals to generate recommendations across servers, communities, and content.
  • Build and deploy machine learning models that drive lifts in new-user retention, combining purpose-built methods with shared foundations such as user and server embeddings.
  • Collaborate cross-functionally to shape the Onboarding roadmap, pairing fast learning experiments with the long-term complexity and scalability of the systems behind them.

What you bring

  • 5+ years of experience in Machine Learning, Data Science, or Applied Science, including a track record shipping recommendation or personalization systems end to end.
  • Hands-on experience with modern recommendation systems, including retrieval and ranking pipelines, embedding-based retrieval such as two-tower models, and deep learning approaches such as transformers and multi-task learning.
  • Strong proficiency in Python and ML frameworks including PyTorch, JAX, or TensorFlow.
  • Experience taking ML systems from ideation to production at scale, including feature pipelines, model serving, monitoring, and online experimentation.
  • Product intuition and a passion for building 0-to-1 ML applications grounded in user feedback and real-world impact.
  • Excellent communication and collaboration skills to represent ML within a cross-functional squad and align stakeholders on trade-offs, priorities, and progress.
  • The ability to thrive in ambiguous environments and tackle open-ended, technically challenging problems.

Technologies

Python, PyTorch, JAX, TensorFlow, transformers, multi-task learning, two-tower models

Benefits

  • Medical, dental, and vision insurance
  • 401(k) plan
  • Life and disability insurance
  • Paid time off
  • Paid parental leave
  • Eligible for equity

Bonus points

  • Experience with cold-start or new-user personalization, onboarding flows, or preference elicitation.
  • Built user representation models or applied social-graph signals to recommendation problems.
  • Worked on growth or engagement problems for consumer products, such as community and content recommendations, user activation, or retention.
  • Experience with large-scale recommendation systems, data infrastructure, distributed computing, and vector search.

Location: San Francisco Bay Area, CA (onsite)

Compensation: USD 220,000 - 275,000 per year

Similar Jobs