Skip to main content

Lyft

ML Software Engineer, ETA

San Francisco, CA$140.8k–$176kfull-timemidAdded today

About this role

Join Lyft's ETA team to build and optimize the machine learning systems that predict ride times across millions of requests daily. You'll develop scalable ML models in production, balancing accuracy with performance while collaborating with engineers and data scientists on a hybrid schedule in San Francisco.

What you'll do

  • Conduct data analysis and build proof-of-concept models to evaluate ML versus non-ML solutions
  • Develop statistical, machine learning, and optimization models for ETA prediction
  • Write production-grade, scalable code serving millions of daily requests
  • Make strategic tradeoffs between model accuracy, complexity, and runtime performance
  • Participate in code reviews, design reviews, and production on-call support
  • Write tested, maintainable, and well-documented code

What they're looking for

  • Machine learning model development (3+ years experience)
  • Python and Go programming
  • Apache Spark and distributed data processing
  • Apache Airflow for pipeline orchestration
  • AWS cloud platform
  • Kubernetes containerization
  • Statistical modeling and optimization
  • Code quality and testing practices

Benefits

  • Medical, dental, and vision insurance with wellness programs
  • Mental health benefits and family building support
  • 401(k) with company match
  • 18 weeks paid parental leave
  • Discretionary PTO for salaried employees; 15 days for hourly
  • Monthly Lyft credits and complimentary Lyft Pink membership
Apply on the employer's site

Opens the official application on the employer’s site. No login required.

Lyft

Lyft develops micromobility hardware solutions including electric bikes and scooters for urban transportation. The company is hiring firmware test engineers and other technical roles to ensure the quality and sustainability of its products.

Website
lyft.com
View all jobs at Lyft

Likely interview questions

  • Walk us through a machine learning project where you had to balance model accuracy against production latency constraints.
  • How have you approached building and deploying models to handle millions of requests per day?