← Projects

Pathwise

AI-powered learning platform that builds a personalized week-by-week roadmap from a 5-question onboarding conversation.

What it is

Pathwise is a learning platform where you describe what you want to learn, and the AI builds a structured, time-tracked roadmap tailored to your goals and availability — not a generic course list.

The experience starts with a short guided conversation: what skill do you want to learn, what's your background, how many hours a day can you commit, and what's your target timeline. From those answers, the platform generates a week-by-week plan with curated resources (YouTube videos, official docs, tutorials) matched to your level.

Why I built it

Every course platform gives you the same content in the same order regardless of who you are. A backend engineer learning machine learning needs a completely different path than a student learning it from scratch. I wanted a tool that actually accounts for that — and that tracks your progress against a realistic timeline, not just a checkbox list.

How it works

The onboarding flow is a 5-step guided chat that collects your goal, background, daily hours, and timeline. That context gets sent to Claude (Opus for path generation, Haiku for supporting tasks) which returns a structured week-by-week plan with resources per milestone.

Progress is tracked daily: you log time spent, mark resources complete, and the platform adjusts your streak and XP. Milestone tests appear automatically at the end of each phase. On-demand tests are available on the paid plan.

The architecture is split across two repos: a Next.js frontend deployed on Vercel and an Express + TypeScript API on Railway. All AI calls use structured outputs so the response is always parseable. Prompt caching is active on system prompts over 1024 tokens.

Key features

  • Guided onboarding chat → personalized roadmap generation
  • Week-by-week plan with curated YouTube + docs resources
  • Daily progress tracking with streak and XP system
  • Milestone badges and path completion rewards
  • Auto-generated tests at key milestones
  • Free tier (1 active roadmap) vs paid (unlimited + on-demand tests)
  • Payments via Lemon Squeezy

Status

In active development. MVP feature scope is defined and architecture is set up. Currently building the core onboarding flow and path generation.