Staff ML Engineer - Genai

Uber Uber · Consumer · Hyderabad, India · Engineering

Staff ML Engineer focused on designing, developing, and productionizing Conversational GenAI solutions for customer support at Uber Eats, aiming for significant cost savings and improved user experience. The role involves agentic AI design, NLP, distillation, experimentation, and technical leadership.

What you'd actually do

  1. Design, develop, and productionize Conversational GenAI solutions in the field of customer support engineering spanning generative AI algorithms, agentic AI design at scale, NLP for query understanding and ranking responses, distillation techniques, etc.
  2. Productionize and deploy these models for real-world applications in customer support.
  3. Design and analyze experiments using a combination of data analysis/statistical analysis to lead the team to a reasonable inference.
  4. Review code and designs of teammates, providing constructive feedback.
  5. Technically lead the team, mentor and guide engineers

Skills

Required

  • Experience in building and owning Conversational GenAI models
  • strong understanding of product and operational metrics
  • Machine learning
  • optimization
  • ML packages such as Tensorflow, PyTorch, JAX, and Scikit-Learn
  • Solid understanding of statistical analysis
  • feature engineering techniques
  • experimental design and analysis
  • exploratory data analysis
  • statistical analysis
  • SQL in a production environment

Nice to have

  • Bachelor's or Master's in Computer Science, Statistics, or a related field or Equivalent Experience in Conversational GenAI
  • dashboarding and using data visualization tools
  • sampling, statistical estimates, descriptive statistics, or similar

What the JD emphasized

  • Conversational GenAI
  • agentic AI design at scale
  • productionize
  • production environment
  • Minimum 10+ years of experience

Other signals

  • designing conversational GenAI systems
  • enhance customer support experience
  • millions of Uber Eats users
  • O(100s millions) cost savings