Lyft is hiring a Machine Learning Engineer in Toronto!
Lyft is hiring a Machine Learning Engineer for its Safety and Customer Care (SCC) team in Toronto. SCC handles more than 1.7 million monthly human and AI interactions and is the company's main direct touchpoint with riders and drivers, so the team's systems sit in a safety-critical, customer-facing environment. The role centres on agentic AI: adapting open-source language models to Lyft's support and safety domain, assembling them into multi-step agents, and building the evaluation infrastructure that demonstrates they are safe enough to ship. The engineer joins a cross-functional group of ML, data, backend and product engineers, data scientists and operations partners. This is a hybrid position requiring at least three days per week in the office, including Mondays, Wednesdays and Thursdays, with up to four weeks per year of remote flexibility.
Key responsibilities
- Post-train and adapt open-source LLMs for SCC use cases using SFT, LoRA and preference-tuning approaches such as RLHF, RLAIF and RLVR.
- Design and ship AI support agents and end-to-end safety case management agents, defining tools, state and control flow in LangGraph or a comparable framework.
- Own the evaluation data flywheel across offline and online settings, including golden sets, rubric-based grading, LLM-as-judge where suitable, regression testing and benchmarks for the team to improve against.
- Convert interaction feedback into training data and learning signals that continuously improve model quality.
- Productionize models and agents at millions of interactions per month, with the monitoring and guardrails needed to operate them safely.
- Apply classical ML methods such as classification, ranking and gradient-boosted trees where they fit better than LLMs.
- Partner with product, operations and data science to scope problems and agree on success metrics.
Key requirements
- 3+ years of industry experience in applied ML or AI, together with an MS or PhD in computer science, machine learning, AI or a related technical field.
- Hands-on post-training experience with open-source models, including fine-tuning and preference-tuning paradigms (SFT, LoRA, RLHF, RLAIF, RLVR).
- Demonstrated agentic development experience, having built and shipped agents through the full development lifecycle.
- Experience designing metrics and offline/online evaluation for generative systems.
- Track record deploying ML or AI applications to real-time production use cases.
- Strong Python programming and practical PyTorch experience.
- Preferred: prior work applying ML/AI to customer support or trust and safety workflows, a PhD, or publications at venues such as NeurIPS, ICML, ICLR, ACL or CVPR.
- Seniority: Mid-level (3+ years of industry experience)
- Work mode: Hybrid — at least 3 days per week in the Toronto office (Mondays, Wednesdays, Thursdays)
- Salary: CAD $118,800 - CAD $148,500 base pay, excluding equity, bonus and benefits
- Key skills: Python, PyTorch, LLM post-training, SFT, LoRA, RLHF, RLAIF, RLVR, LangGraph, agentic AI, LLM evaluation, LLM-as-judge, benchmarking, production ML deployment, monitoring and guardrails
Want more jobs?
Join our mailing list to receive updates about new data science and Machine Learning Engineer jobs.
