2 months ago
Toronto, CanadaMid Level
Responsibilities
- Conduct literature reviews and build post-training frameworks and lifecycles for open-source models
- Curate and process human and synthetic data for SFT, LoRA, RLHF, RLAIF, and RLVR
- Develop, evaluate, and productionize AI agents using LangGraph or equivalent frameworks
- Design agent tools, state, and control flow across the full agent development lifecycle
- Build evaluation frameworks, golden sets, rubric-based grading, LLM-as-judge systems, and regression tests
- Ship models and agents to real-time production with monitoring and safety guardrails
- Apply classification, ranking, and gradient-boosted trees when appropriate
- Partner with product, operations, and data science teams to define problems and success metrics
- Convert interaction feedback into training data and learning signals to continuously improve models
Requirements
- At least three years of industry experience in applied ML/AI, inclusive of an MS or PhD in a related technical field
- Post-training experience with open-source models, including SFT, LoRA, RLHF, RLAIF, and RLVR
- Experience building and shipping agents with LangGraph or equivalent agentic frameworks
- Experience designing metrics and offline and online evaluations for generative AI systems
- Experience deploying ML/AI applications to real-time production environments
- Strong programming skills in Python and hands-on experience with PyTorch
- Preferred experience applying ML/AI to customer support or trust and safety workflows
- A PhD in Computer Science, Machine Learning, Statistics, or a related technical field is preferred
- Preferred publications at top-tier venues such as NeurIPS, ICML, ICLR, ACL, or CVPR
Benefits
- Extended health and dental coverage, life insurance, and disability benefits
- Mental health, family building, child care, and pet benefits
- Lyft-funded Health Care Savings Account
- RRSP plan with company match
- Flexible paid time off for salaried employees or 15 paid days off for hourly employees, plus additional annual accrual
- 18 weeks of paid parental leave for biological, adoptive, and foster parents
- Subsidized commuter benefits and Lyft ride credits
- Hybrid in-office schedule requiring at least three days per week on Mondays, Wednesdays, and Thursdays, with up to four weeks per year working from anywhere
Categories
About Lyft
Lyft is a NASDAQ-listed transportation marketplace that connects riders with drivers and offers bike and scooter rentals in one app. Founded in 2012 and headquartered in San Francisco, it operates across the U.S. and parts of Canada. Its revenue comes from service fees on rides and rentals, enterprise programs, and consumer subscriptions such as Lyft Pink.