Data Scientist (ML Engineer)

Adelphi Group•Toronto, ON

8d•CA$120,000 - CA$140,000•Hybrid

About The Position

As a Data Scientist (ML Engineer), on the Perpetua team, you will design, experiment with, and ship the machine learning systems that decide how thousands of brands spend their advertising budgets across retail media. This is the engine that takes autonomous action on the customer’s behalf. It is not a model that produces recommendations for someone else to act on, but the system that sets bids and allocates spend in production, in real time, against each advertiser’s goals. Your work runs live across thousands of customers worldwide. Our team primarily works with Python and the Google Cloud Platform suite of products like Cloud Run and Vertex AI to productize cutting-edge data features. We are currently working on developing a scalable advertising bidding platform that enables advertisers to implement custom and versatile bidding strategies including but not restricted to maximizing advertising sales, dominating top-of-search placements, optimizing for total sales, incremental sales, new-to-brand purchases, organic rank, etc. Increasingly, this work sits alongside a newer layer of generative and agentic AI; LLM-based reasoning that plans, explains, and reacts to natural language goals. Knowing where classical optimization is the right tool and where the generative layer adds leverage is part of the craft on this team.

Requirements

5+ years of experience as a data scientist or engineer working with data scientists
Strong experience with algorithms and data pipelines processing terabytes of data per day
You have experience taking concepts from inception through to production and ongoing monitoring and enhancements
Experience in retail media, digital advertising, or e-commerce is an asset
You have worked in organizations with cross-functional teams of ~5 people, solving hard problems collaboratively and working tightly with your immediate team members and across the organization
Working knowledge of reinforcement learning and linear/non-linear optimization is a strong asset, given how central these techniques are to the bidding engine
Curiosity about applied LLMs and agentic systems, and comfort using modern AI-assisted development tools (such as Claude Code) as part of how you build
Able to create and make changes to traditional ML models, including but not limited to Linear regression, XGBoost and Logistic regression
Competent in training and evaluating models using mainstream data science tools including but not limited to sklearn, Pandas, keras and/or PyTorch
Experience in cloud native ML training platforms like BigQuery ML or Snowpark

Responsibilities

Work across retail media (starting with Amazon) to understand the intricate relationships between bids, placement, conversion, and sales, and turn that understanding into systems that optimize advertising autonomously on the customer’s behalf.
Design, Implement, and Analyze experiments for deriving Actionable Insights.
Analyze advertising performance data to improve the core strategies that power Perpetua's advertising engine.
Help define how Perpetua’s machine-learning optimization works alongside the emerging generative and agentic layer, deciding where reinforcement learning and classical optimization are the right tools, and where LLM-based reasoning meaningfully improves how the platform plans and explains its decisions.
Support the growth of the team by contributing to activities for establishing best practices, recruitment, and authoring design documents.