All projects

MIT Sloan · Product-management course project

Helping student teams turn scattered course requirements into a shared plan, starting with AI syllabus extraction.

My role
Product lead, deployment & landing site
When
2026
Stage
Completed MVP & pilot
Duration
2-week MVP build
Team
4 people

The key decision

Prioritize syllabus deadline extraction over scheduling so a team gets value before it has to maintain another workspace.

The evidence

An eight-user pilot reported 100% task completion and a 30% reduction in coordination time.

What I built

I deployed the React and Express app to Vercel and Supabase, resolved API bundling and PDF-parser compatibility issues, fixed proxy-dependent login sessions and silent analysis timeouts, and built the public landing site. A teammate built the core app.

8users in the pilot
100%task completion in the pilot
30%reported reduction in coordination time
Aligned

The problem: the work around the work

Student teams manage deadlines in syllabi, decisions in chat, and documents across several tools. Someone becomes the unofficial coordinator. The challenge is making those pieces usable together without creating another workspace to maintain.

My role

For a MIT Sloan product-management course project, I interviewed MIT Sloan and Harvard graduate students, defined a seven-feature roadmap, and prioritized the first usable workflow.

A teammate built the core app. I owned getting it running in production on Vercel and Supabase, and built the public landing site. The MVP was built and deployed in two weeks.

The key trade-off

I prioritized AI syllabus deadline extraction over scheduling. A syllabus already contains information every team needs, so turning it into a shared timeline creates value at the start of a project. Scheduling depends on more setup and participation before the team sees a benefit.

The product brings existing documents and communication channels into one project view. A team can begin with a syllabus and invite link, then use additional features as needed.

From syllabus to shared plan

A team member uploads a syllabus. AI extracts deadlines, deliverables, and grading weights into a shared timeline. Linked documents and channels keep the source material findable, while shared action items support follow-through.

What I built: from prototype to production

Aligned is a TypeScript full-stack app: React and Vite on the front end, an Express API, PostgreSQL on Supabase through Drizzle ORM, and Claude for syllabus extraction, with live updates over Server-Sent Events. The app started in Replit, where it ran as one long-lived server.

Moving that server into Vercel serverless functions surfaced problems local development hid. I bundled the API with esbuild and resolved an ESM and CommonJS conflict, pinned a PDF parser version that did not depend on browser-only APIs, created the session table in Supabase at startup, and fixed login sessions that broke behind Vercel's proxy.

I also fixed a syllabus-analysis timeout that failed silently and email notifications that were cut off when a function returned before the send finished. I then built the landing site with a product demo and routed the app subdomain straight to login.

Pilot results

The eight-user pilot reported 100% task completion and a 30% reduction in coordination time. These are results from an early, small pilot rather than evidence of sustained adoption.

The strongest product lesson was to make the first useful action happen before asking users to reorganize how they work.

Current status

Aligned is a completed course MVP and pilot. The landing page and source code are public, and the demo video below shows the full workflow from syllabus upload to shared timeline.

Project materials

Aligned in action: AI syllabus analysis, shared timeline, and real-time action items