Learn to Earn
A guided learn-to-earn platform that collapses the gap between learning a skill and earning with it: instead of finishing courses with nothing to show, users deliver real projects step by step with AI guidance.
Status :
In active development
Stack :
Next.js 16 · React 19 · Prisma · Neon · Anthropic API
My Role :
Product Designer & Builder (vibecoded)
Client :
Self-initiated
Project Health
31
Data models powering the platform
16
Product phases shipped
1
Guided learn-to-earn loop
Overview
Learn to Earn is a platform I designed and vibecoded to collapse the longest, most discouraging gap in any skill journey: the distance between learning something and earning from it. Instead of tutorials followed by a cold job hunt, users learn by completing real, guided projects, delivering finished work while being coached through every step.
The dream: cut the learn-to-earn timeline from months of theory to a guided sprint that ends with something delivered, and a skill proven by the delivery.
The Problem
The conventional path is learn, then hope. People spend months on courses that end with a certificate and no output, then start from zero trying to convert that into income. The theory rarely survives contact with a real deliverable, and the gap between finishing a course and getting paid is where most learners stall out.
Objectives
Replace passive tutorials with guided delivery of real projects.
Give every learner a finished, portfolio-grade output as the default result, not a lucky exception.
Shorten time-to-competence by coaching through the actual work instead of around it.
Make progress legible, so learners always know how far into a project and a skill they are.
How It Works
I built the platform end to end as a modern web app: Next.js 16, React 19, Prisma and Neon Postgres, with an AI coaching layer on the Anthropic API. The core is a project engine that breaks real work into phases and steps with dependencies, tracks weighted completion, and guides the learner forward one unblocked step at a time.
Key Decisions
1. Guided delivery over passive tutorials
The whole product is organized around finishing a real project, not watching lessons. Every learner works toward a delivered artifact, and the coaching happens inside the work. That single choice is what compresses the learn-to-earn timeline.
2. Weighted progress with phase dependencies
Projects decompose into phases and steps with real dependencies and cycle detection, and progress is weighted so the bar reflects genuine advancement, not step counts. Learners always see exactly where they are and what unblocks next.
3. An AI layer that plans, not just chats
The AI generates structured learning plans and decomposes feedback into actionable steps, validated against typed schemas. It behaves like a project guide with a checklist, not a chatbot, so guidance is always tied to the next concrete move.
The Solution
A project engine: phases, steps and dependencies with cycle detection and weighted progress.
AI coaching: structured plan generation and feedback decomposition into next actions.
Delivery tooling: file handling with signed, expiring URLs so work is submitted and reviewed in-platform.
A dashboard tying learning plans, active projects, completion and earnings into one view.
Results
The platform is substantially built: sixteen product phases shipped across a 31-model schema, including auth, the project engine, the AI planning and feedback layer, file handling, dashboard and an earnings/completion flow. What remains is a resource-verification pipeline and hardening, which is why completion sits at 85% rather than done.
Reflections
What works: guided delivery. Coaching inside a real project, rather than before one, is the mechanism that actually shortens the path to earning.
What I would improve: the resource search and verification pipeline is the missing piece to make plans fully self-serve.
What I would validate next: real time-to-first-delivery for new users, the number the whole product exists to move.
The AI generated build plan, grouped into phases with dependencies, time estimates and a step by step checklist.
Project overview: scope, deliverables, constraints and tools captured up front so the plan has something to reason about.
A learning plan broken into modules and tasks, tracked to completion alongside the project it supports.
The create flow, where a plain language goal becomes either a project, a learning plan or a quick Blitz answer.
Projects list with status and completion at a glance.
Learning plans list, separating active work from drafts.
Skills record, built automatically from completed work rather than self declared.
The same dashboard on mobile, since most check ins happen away from the desk.
CLIENT PROOF
Nick Simmons
founder of JAH creative design agency
Having Lawrence work under my agency made me money as an efficient designer he is, helped me retain my status as a Top 4% upwork agency and I can't be any less proud of him. Under me he learnt framer just so we could get more jobs and he delivered.
Loong Pancoast
CEO of DriveOn
Lawrence is a professional through and through. He has been an absolute pleasure to work with, taking in account all of our needs while meeting expectations. I would 100 percent recommend Lawrence for any developmental design needs. If you are hesitating, don't, you won't regret it.
Drew Levin
Co-founder of SideShift
Lawrence worked hand in hand with me and Nick Lawton to get the first version of our site out the door, staying flexible through every change we needed. His cooperation is a big reason we could launch on time and begin major operations. A designer who listens closely and delivers.
12:27
Learn to Earn
A guided learn-to-earn platform that collapses the gap between learning a skill and earning with it: instead of finishing courses with nothing to show, users deliver real projects step by step with AI guidance.
Status :
In active development
Stack :
Next.js 16 · React 19 · Prisma · Neon · Anthropic API
My Role :
Product Designer & Builder (vibecoded)
Client :
Self-initiated
Project Health
31
Data models powering the platform
16
Product phases shipped
1
Guided learn-to-earn loop
Overview
Learn to Earn is a platform I designed and vibecoded to collapse the longest, most discouraging gap in any skill journey: the distance between learning something and earning from it. Instead of tutorials followed by a cold job hunt, users learn by completing real, guided projects, delivering finished work while being coached through every step.
The dream: cut the learn-to-earn timeline from months of theory to a guided sprint that ends with something delivered, and a skill proven by the delivery.
The Problem
The conventional path is learn, then hope. People spend months on courses that end with a certificate and no output, then start from zero trying to convert that into income. The theory rarely survives contact with a real deliverable, and the gap between finishing a course and getting paid is where most learners stall out.
Objectives
Replace passive tutorials with guided delivery of real projects.
Give every learner a finished, portfolio-grade output as the default result, not a lucky exception.
Shorten time-to-competence by coaching through the actual work instead of around it.
Make progress legible, so learners always know how far into a project and a skill they are.
How It Works
I built the platform end to end as a modern web app: Next.js 16, React 19, Prisma and Neon Postgres, with an AI coaching layer on the Anthropic API. The core is a project engine that breaks real work into phases and steps with dependencies, tracks weighted completion, and guides the learner forward one unblocked step at a time.
Key Decisions
1. Guided delivery over passive tutorials
The whole product is organized around finishing a real project, not watching lessons. Every learner works toward a delivered artifact, and the coaching happens inside the work. That single choice is what compresses the learn-to-earn timeline.
2. Weighted progress with phase dependencies
Projects decompose into phases and steps with real dependencies and cycle detection, and progress is weighted so the bar reflects genuine advancement, not step counts. Learners always see exactly where they are and what unblocks next.
3. An AI layer that plans, not just chats
The AI generates structured learning plans and decomposes feedback into actionable steps, validated against typed schemas. It behaves like a project guide with a checklist, not a chatbot, so guidance is always tied to the next concrete move.
The Solution
A project engine: phases, steps and dependencies with cycle detection and weighted progress.
AI coaching: structured plan generation and feedback decomposition into next actions.
Delivery tooling: file handling with signed, expiring URLs so work is submitted and reviewed in-platform.
A dashboard tying learning plans, active projects, completion and earnings into one view.
Results
The platform is substantially built: sixteen product phases shipped across a 31-model schema, including auth, the project engine, the AI planning and feedback layer, file handling, dashboard and an earnings/completion flow. What remains is a resource-verification pipeline and hardening, which is why completion sits at 85% rather than done.
Reflections
What works: guided delivery. Coaching inside a real project, rather than before one, is the mechanism that actually shortens the path to earning.
What I would improve: the resource search and verification pipeline is the missing piece to make plans fully self-serve.
What I would validate next: real time-to-first-delivery for new users, the number the whole product exists to move.
The AI generated build plan, grouped into phases with dependencies, time estimates and a step by step checklist.
Project overview: scope, deliverables, constraints and tools captured up front so the plan has something to reason about.
A learning plan broken into modules and tasks, tracked to completion alongside the project it supports.
The create flow, where a plain language goal becomes either a project, a learning plan or a quick Blitz answer.
Projects list with status and completion at a glance.
Learning plans list, separating active work from drafts.
Skills record, built automatically from completed work rather than self declared.
The same dashboard on mobile, since most check ins happen away from the desk.
CLIENT PROOF
Nick Simmons
founder of JAH creative design agency
Having Lawrence work under my agency made me money as an efficient designer he is, helped me retain my status as a Top 4% upwork agency and I can't be any less proud of him. Under me he learnt framer just so we could get more jobs and he delivered.
Loong Pancoast
CEO of DriveOn
Lawrence is a professional through and through. He has been an absolute pleasure to work with, taking in account all of our needs while meeting expectations. I would 100 percent recommend Lawrence for any developmental design needs. If you are hesitating, don't, you won't regret it.
Drew Levin
Co-founder of SideShift
Lawrence worked hand in hand with me and Nick Lawton to get the first version of our site out the door, staying flexible through every change we needed. His cooperation is a big reason we could launch on time and begin major operations. A designer who listens closely and delivers.
12:27
Learn to Earn
A guided learn-to-earn platform that collapses the gap between learning a skill and earning with it: instead of finishing courses with nothing to show, users deliver real projects step by step with AI guidance.
Status :
In active development
Stack :
Next.js 16 · React 19 · Prisma · Neon · Anthropic API
My Role :
Product Designer & Builder (vibecoded)
Client :
Self-initiated
Project Health
31
Data models powering the platform
16
Product phases shipped
1
Guided learn-to-earn loop
Overview
Learn to Earn is a platform I designed and vibecoded to collapse the longest, most discouraging gap in any skill journey: the distance between learning something and earning from it. Instead of tutorials followed by a cold job hunt, users learn by completing real, guided projects, delivering finished work while being coached through every step.
The dream: cut the learn-to-earn timeline from months of theory to a guided sprint that ends with something delivered, and a skill proven by the delivery.
The Problem
The conventional path is learn, then hope. People spend months on courses that end with a certificate and no output, then start from zero trying to convert that into income. The theory rarely survives contact with a real deliverable, and the gap between finishing a course and getting paid is where most learners stall out.
Objectives
Replace passive tutorials with guided delivery of real projects.
Give every learner a finished, portfolio-grade output as the default result, not a lucky exception.
Shorten time-to-competence by coaching through the actual work instead of around it.
Make progress legible, so learners always know how far into a project and a skill they are.
How It Works
I built the platform end to end as a modern web app: Next.js 16, React 19, Prisma and Neon Postgres, with an AI coaching layer on the Anthropic API. The core is a project engine that breaks real work into phases and steps with dependencies, tracks weighted completion, and guides the learner forward one unblocked step at a time.
Key Decisions
1. Guided delivery over passive tutorials
The whole product is organized around finishing a real project, not watching lessons. Every learner works toward a delivered artifact, and the coaching happens inside the work. That single choice is what compresses the learn-to-earn timeline.
2. Weighted progress with phase dependencies
Projects decompose into phases and steps with real dependencies and cycle detection, and progress is weighted so the bar reflects genuine advancement, not step counts. Learners always see exactly where they are and what unblocks next.
3. An AI layer that plans, not just chats
The AI generates structured learning plans and decomposes feedback into actionable steps, validated against typed schemas. It behaves like a project guide with a checklist, not a chatbot, so guidance is always tied to the next concrete move.
The Solution
A project engine: phases, steps and dependencies with cycle detection and weighted progress.
AI coaching: structured plan generation and feedback decomposition into next actions.
Delivery tooling: file handling with signed, expiring URLs so work is submitted and reviewed in-platform.
A dashboard tying learning plans, active projects, completion and earnings into one view.
Results
The platform is substantially built: sixteen product phases shipped across a 31-model schema, including auth, the project engine, the AI planning and feedback layer, file handling, dashboard and an earnings/completion flow. What remains is a resource-verification pipeline and hardening, which is why completion sits at 85% rather than done.
Reflections
What works: guided delivery. Coaching inside a real project, rather than before one, is the mechanism that actually shortens the path to earning.
What I would improve: the resource search and verification pipeline is the missing piece to make plans fully self-serve.
What I would validate next: real time-to-first-delivery for new users, the number the whole product exists to move.
The AI generated build plan, grouped into phases with dependencies, time estimates and a step by step checklist.
Project overview: scope, deliverables, constraints and tools captured up front so the plan has something to reason about.
A learning plan broken into modules and tasks, tracked to completion alongside the project it supports.
The create flow, where a plain language goal becomes either a project, a learning plan or a quick Blitz answer.
Projects list with status and completion at a glance.
Learning plans list, separating active work from drafts.
Skills record, built automatically from completed work rather than self declared.
The same dashboard on mobile, since most check ins happen away from the desk.
CLIENT PROOF
Nick Simmons
founder of JAH creative design agency
Having Lawrence work under my agency made me money as an efficient designer he is, helped me retain my status as a Top 4% upwork agency and I can't be any less proud of him. Under me he learnt framer just so we could get more jobs and he delivered.
Loong Pancoast
CEO of DriveOn
Lawrence is a professional through and through. He has been an absolute pleasure to work with, taking in account all of our needs while meeting expectations. I would 100 percent recommend Lawrence for any developmental design needs. If you are hesitating, don't, you won't regret it.
Drew Levin
Co-founder of SideShift
Lawrence worked hand in hand with me and Nick Lawton to get the first version of our site out the door, staying flexible through every change we needed. His cooperation is a big reason we could launch on time and begin major operations. A designer who listens closely and delivers.
12:27