About this role
Job Description Role Overview:
The Government of Ontario is embarking on a large Data & AI transformation initiative. Core to the transformation is the enablement of the Ontario Data Platform (ODP), which is being developed as the enterprise technology ecosystem for everything Data & AI. ODP is a modern ecosystem of data solution consisting of data ingestion technologies, storage, processing and visualization as well as solutions related to data governance, quality, marketplace, etc. ODP will initially serve all internal government clusters, ministries and program areas, but also external agencies in the future. We are looking for a data scientist to support various initiatives related to the platform as well as the execution of use cases to support the needs of various businesses within the government.
Key Responsibilities:
• Run and Monitor Inferences:
• Execute batch inference jobs on Databricks and other environments.
• Troubleshoot inference issues, log results, and escalate model anomalies.
• Pipeline & Evaluation Support:
• Assist with running DSPy pipelines and loading custom models/configs.
• Support evaluation runs using internal metrics
• Track and summarize performance metrics for inference experiments.
• Data Handling & Experimentation:
• Help prepare and clean legal text datasets for inference and evaluation.
• Annotate outputs to support error analysis and benchmarking.
• Maintain reproducible scripts/notebooks for inference runs.
Responsibilities:
• Lead the development and delivery of functional and ministry-specific analytics to support evidence-based decision-making and produce actionable insights.
• Work closely with client?s groups to assess current data analytics and reporting capabilities, gather future-state requirements and identify further opportunities for improvement.
• Facilitate decision-making and manage client expectations.
• Own the execution of analytics initiatives including end-to-end reporting and data set delivery.
• Develop robust statistical models and machine learning algorithms to model business scenarios and extract valid inferences. <li class="align-justify" style="text-align:ju