Senior AI Systems Engineer — Production LLM Pipelines

Urban Ridge Supplies; Toronto; Hybrid; CAD 90,000 - 133,000; Full time

Job summary

Morningstar is seeking a Senior Software Engineer to build applied systems for AI capabilities in production, including data pipelines, LLM integrations, agentic workflows, and evaluation frameworks. This role focuses on applied engineering, delivering durable architectures and swappable components to adopt new models with evidence.

We offer a hybrid work environment with in-person collaboration weekly and four days in-office in most locations, plus a range of benefits to support flexibility and

Qualifications

• 5+ years of software engineering with production API and data-pipeline design.
• Hands-on experience integrating LLMs and eval tooling.
• Strong Python development skills across services and APIs.

Responsibilities

• Integrate with hosted LLM inference and evaluate models for cost/performance.
• Design adapters to sync source systems into a governed dataset.
• Build and maintain an eval harness to catch regressions before release.
• Design MCP tools and API/GraphQL interfaces separating fuzzy retrieval from exact lookups.
• Operate pipelines for governance and versioning of entities in serving layer.
• Develop LLM-driven extraction and classification agents for human review.
• Own structured content modeling and API surface versions with other teams.

Skills

Python

TypeScript/Node.js

REST API

GraphQL

Data pipelines

AWS

Tools

FastAPI

GraphQL

AWS

Kubernetes

Job description

Morningstar is seeking a Senior Software Engineer to build applied systems for AI capabilities in production, including data pipelines, LLM integrations, agentic workflows, and evaluation frameworks. This role focuses on applied engineering, delivering durable architectures and swappable components to adopt new models with evidence.

We offer a hybrid work environment with in-person collaboration weekly and four days in-office in most locations, plus a range of benefits to support flexibility and