Skip to content

Work

Case studies.

Product and engineering decisions, with the evidence behind them.

Translate scanned documents into editable Google Docs, preserving tables, headings and reading direction.

Verto: AI Translation Platform

Product type
SaaS
Capability
Document AI
Role
Product Engineer
Build
Built in 3 months

The challenge

Scanned documents meant translating the text, reconstructing the layout and preparing a deliverable, then managing the customer, job history and payment around it.

What I built

I designed Verto in Figma and built the entire platform, from document processing and editing to customer management, billing, administration and deployment.

  • Next.js
  • OpenAI
  • Stripe
  • Google Docs API
  • TypeScript
  • Supabase
  • PostgreSQL
  • Trigger.dev
  • Tiptap
Next.js
OpenAI
Stripe
Google Docs APITypeScriptSupabasePostgreSQLTrigger.devTiptap
TypeScriptSupabasePostgreSQLTrigger.devTiptapGoogle Docs API
Help study-abroad advisors find suitable programs, understand each match and build a student shortlist.

University Program Matching

Product type
Internal tool
Capability
AI matching
Role
Product Engineer
Build
About 14 weeks · 2025

The challenge

Four hundred students were waiting for program recommendations. Keyword matching missed related subjects, while admission rules were buried in inconsistent spreadsheets.

What I built

I designed and built the complete matching platform, connecting the academic catalog, AI import pipeline and explainable recommendations in a workspace advisors could use during a student call.

  • Laravel
  • OpenAI
  • MeiliSearch
  • PHP
  • Livewire
  • Alpine.js
  • MySQL
  • Redis
Laravel
OpenAI
MeiliSearch
PHPLivewireAlpine.jsMySQLRedis
LivewireAlpine.jsMySQLRedisPHP
Replace WhatsApp document collection with guided student applications that feed the advisor's existing records.

Student Applications & Document Collection

Product type
Client portal
Capability
Document automation
Role
Product Engineer
Build
Built in 6 weeks

The challenge

Applications arrived as WhatsApp messages: passports here, transcripts elsewhere. Staff chased missing files and retyped details already sitting on the documents.

What I built

I designed and built a four-language student portal: apply before every document is ready, review extracted details and resume across devices, with files and grades reaching the agency’s CRM.

  • Next.js
  • Laravel
  • AWS Textract
  • React
  • TypeScript
  • PHP
  • OpenAI
  • S3
  • next-intl
  • TanStack Query
Next.js
Laravel
AWS Textract
ReactTypeScriptPHPOpenAIS3next-intlTanStack Query
TypeScriptPHPOpenAIS3next-intlTanStack QueryReact

Agency CRM & Billing

Product type
CRM / ERP
Capability
Billing
Role
Product Engineer
Platform evolution
Ongoing development

The challenge

Every student brought an application, documents to collect and money to track. Staff needed one reliable view of the case and its outstanding balance.

What I built

I designed and built the agency's CRM and billing platform, linking customer records, staff access, invoices and payments so the portal and matching system could share them.

  • Laravel
  • Cash Plus
  • MySQL
  • PHP
  • Livewire
  • Alpine.js
  • Redis
  • MeiliSearch
Laravel
Cash Plus
MySQL
PHPLivewireAlpine.jsRedisMeiliSearch
LivewireAlpine.jsRedisMeiliSearchPHP

Confidential work

Some client work stays private. I can discuss the problems, my contribution and the engineering approach without sharing confidential material. Ask about relevant work.

Have a project in mind?

Discuss a projectDiscuss a projectDiscuss a projectDiscuss a project
© 2026 Yassine Chettouchhello@yaschet.com
LinkedInGitHubX