Shirley XueAll work
AlefAuthor ToolEdTech2020 – 2022

Lesson templates that couldn't break live classes

Problem
Lesson structure couldn't flex as markets expanded. New templates risked breaking what students already used.
Solution
Four composable blocks with contracts: authors configure the lesson; the apps students and teachers use stay stable.
Result
Shipped to production on a four-block contract, stress-tested against UAE Ministry of Education and Indonesia rollout flows before build. The cross-team design review I started outlived the project.

Role & team

Led design of the authoring and admin experience

  • With product owners and engineering; supported the junior designers on the team.

Duration & scope

2020–2022, four composable blocks

  • The authoring system behind lesson templates across expanding markets.

The constraint

Live classes could not break

  • Students were already inside the lessons this system authored, so a breaking change would land in a live class.
Context

Alef, an adaptive-learning EdTech company, was expanding into new markets faster than its lesson structure could flex, and authors needed to compose lessons without engineering help.

Blocks

The architecture of lesson building blocks

We audited existing lesson structures and ran stakeholder interviews until the complexity reduced to four composable units: the information architecture of every lesson the system could author. Each block carried a clear contract: what authors could configure, what data consumer apps would receive, and what could never be mixed.

Modes

Video, HTML, and rich media. Passive delivery — no learner actions tracked.

LESSONContentAssessmentAssignmentKey Termblocks compose; the frame never changesContent · contractTAKESvideo · HTML · rich mediaGIVES BACKnothing tracked
Patterns

Translating legacy into config

Every legacy structure was reverse-engineered into generic patterns that authors could reason about and developers could implement once. That let us scope the project without guessing: authors saw familiar building blocks, engineering saw a finite set of configurations.

Modes

The default authoring surface: outline, canvas and inspector in one layout.

STANDARD TEMPLATE VIEWOutlineLesson canvasContentAssessmentAssignmentInspectorone layout, whatever the lesson is made of
Flows

Stress-testing regional branch logic

Markets don't share one curriculum logic. Before build, we mapped author configurations against real regional flows: UAE Ministry of Education (MOE) requirements, Indonesia rollout paths, and edge cases from simple linear lessons to asymmetric branching.

Modes

Templates aligned to Ministry of Education curriculum — compliance as a configuration constraint, not a custom build.

ContentMOE gateAssessmentReportCurriculum specMinistry-of-Education templates: compliance sits in the flow as a gate, not as a separate build.
Screens

The shipped authoring surface

The View Template flow runs Big Idea → Depth of Knowledge → Check My Understanding, each step with release conditions.

Alef Author production screens — template library with the View Template branching flow, and the Add Metadata wizard stepView full size
Production screens · template library + View Template flow · Add Metadata step
Design review

Design review and the design system

In parallel I was a key contributor to the Alef design system: I built the proof-of-concept with the dev team and documented the design principles with the lead UI designer.

I also initiated the cross-team design review, where designers from different product teams critiqued each other's work on a regular schedule. It outlived the project.

Landed

Where it landed

The authoring surface shipped to production. There is no outcome metric from this period.