Technical Product Manager, AI Systems & Architecture

About my Blueprint
my Blueprint is a leading developer of K-12 EdTech tools used by over 1 million students across Canada. We create best-in-class digital portfolios and career/life planning tools to support student success and documentation of learning. You can learn more about us at

About the Opportunity
We are building the next generation of student success technology — powered by a unified intelligence layer that connects student work, learning pathways, and educator workflows across my Blueprint and Spaces

EDU. We’re hiring a  Technical Product Manager, AI Systems & Architecture , to define and operationalize the architectural backbone of our AI systems. This role is architect-first, feature-second. You will define the systems that allow AI to understand student work, interpret progress, and support educators — all from one shared platform layer. Your focus is on defining how the system behaves: the logic, flow, and structure that make the experience work.

You will have support from Engineering to implement the underlying infrastructure. You’ll start as one half of a two-person team, initially on your own, with a dedicated AI-focused engineer to be hired shortly after you begin — allowing you to move quickly, think holistically, and deliver early vertical slices that evolve into core platform features.

Success in this role is measured by:

Clear, structured architectural decisions

Crisp documentation and scoping

Systems that are ready for engineering execution

Vertical slices shipped with defined evaluation criteria

This is not a “prompt ideation” or research-heavy strategy role. It is a  systems design and delivery role . If you're energized by technical systems, data architecture, AI reasoning patterns, and designing the backbone of an AI-first product ecosystem, this is the role for you!

What You Will Own
AI Platform Architecture

You will lead the definition of the architectural backbone that enables AI capabilities across the product, including:

Designing the foundational learner data model that AI systems reason over

Defining relationships across students, plans, goals, competencies, assessments, and artifacts

Specifying schemas, metadata strategies, and signal hierarchies

Determining which signals are precomputed vs. generated on-demand

Establishing the system contracts engineers build against

Your work becomes the structural layer that all AI-powered experiences rely on.

Extraction & Retrieval System Design

You will define how raw student data becomes structured intelligence. This includes:

Designing the transformation pipeline from content (text, images, documents) into usable signals

Defining tagging, enrichment, and indexing strategies

Specifying retrieval logic and ranking approaches

Determining how context is assembled for AI reasoning

Defining failure states and fallback behaviours

You will not implement pipelines — but you will define their required behaviour with precision.

Orchestration & Model Routing

You will define the rules that govern how AI models are used across the platform. This includes:

When retrieval is used vs. when generative reasoning is invoked

How different tools and model types interact

Model selection strategies based on cost, latency, and quality constraints

Guardrails for safe and consistent outputs

Shared foundations across AI experiences

You define the playbook that engineers execute.

Cost, Reliability & Guardrails

You will ensure the AI platform scales responsibly. This includes:

Establishing an SLM-first philosophy where appropriate

Defining cost envelopes and routing constraints

Identifying caching and efficiency strategies

Designing monitoring and evaluation signals

Ensuring multi-tenant data isolation and privacy

You ensure AI remains reliable, affordable, and scalable.

Vertical Slice Delivery

You will scope and deliver the first end-to-end AI experiences that demonstrate platform value. This includes:

Defining realistic, constrained vertical slices

Writing clear specifications that move from input → system logic → output → evaluation

Partnering with engineering to ship working demos

Defining quality benchmarks before release

Success is measured by working slices — not abstract strategy.

Inn…