Personal Product

Vocabulary Builder (Personal Product)

A cross-platform vocabulary learning app deployed on web and android, built with TypeScript, Expo and Domain Driven Design. Features include spaced repetition reviews (SM-2), hands-free podcast narration, bilingual (EN/FR) word tracking with definitions and context sentences, and offline-first backup with ZIP-based sync, all from a single codebase with zero cloud dependency.

Timeline
Jan 2026 – Jul 2026
Stack
React Native, Expo

GitHub

Vocabulary Builder store feature graphic
Add words with context

Add words with context

  • Captures the core word-entry workflow for adding vocabulary with definitions and examples.
  • Supports bilingual EN/FR tracking from the same form-driven experience.
  • Stores words locally through the shared domain model and repository layer.
Word of the day

Word of the day

  • Highlights a daily vocabulary prompt to make returning to the app feel useful immediately.
  • Surfaces definitions, translations, and context in a compact mobile-first view.
  • Reinforces regular study habits without requiring a full review session.
Review your words

Review your words

  • SM-2 scheduling prioritizes words that are due instead of showing a flat flashcard queue.
  • Review flow tracks recall quality and updates the next review date automatically.
  • Built to support daily retention with minimal manual organization.
Listen mode

Listen mode

  • Hands-free podcast narration lets users review vocabulary while walking, commuting, or resting.
  • Reads saved words with definitions and context so review is not limited to screen time.
  • Reuses the same vocabulary data as the visual review flow.
Learning statistics

Learning statistics

  • Progress views summarize review activity, streaks, and vocabulary growth.
  • Gives learners feedback on consistency and retention over time.
  • Keeps personal analytics local with the rest of the offline-first app data.
Source-based organization

Source-based organization

  • Words can be grouped by source so learners remember where vocabulary came from.
  • Supports books, articles, videos, courses, or any custom learning context.
  • Source metadata is part of the clean domain model instead of UI-only tagging.
Offline-first backup and preferences

Offline-first backup and preferences

  • Backup and restore workflows use ZIP-based sync to avoid cloud dependency.
  • Language and app preferences are available from a focused settings screen.
  • Native and web storage implementations stay interchangeable behind shared interfaces.
Choose your language

Choose your language

  • Onboarding supports English and French usage from the start.
  • Language choice connects to the app's localized vocabulary workflow.
  • Keeps the first-run experience aligned with the bilingual product scope.

Implementation notes

  • Built and shipped this as a personal product to solve my own vocabulary learning workflow, taking it from idea to production on both Web and Android solo
  • Deployed on Web and Android from a single TypeScript codebase by applying DDD and clean architecture, swapping SQLite (native) and IndexedDB (web) persistence behind a shared repository interface
  • Implemented SM-2 spaced repetition algorithm scheduling daily reviews and a podcast narration mode for hands-free passive review, maximizing retention across multiple study contexts
  • Built offline-first backup system with ZIP compression, incremental push/pull sync, and image bundling eliminating cloud dependency and server costs entirely

Featured Projects

  • Open Source

    8086 Processor Web Emulator

    Open-source Intel 8086 emulator and assembler for the browser (Vanilla JS), built as a second year university project in a team of 6, featuring instruction set simulation and interactive debugging.

    • JavaScript
    • HTML
    • CSS
  • Thesis Project

    IMRaD Analyzer: Distributed Microservices for Scientific Text Classification

    An 8-node microservices platform leveraging BERT and Gemini Pro to classify scientific papers, using a custom 169k-sentence dataset to achieve a 98.21% peak F1-score.

    • Next.js
    • FastAPI
    • BERT
    • Gemini Pro
    • TF Serving

Let's work together

I'm currently open to new roles, remote-first or international. If you're building something ambitious, I'd love to hear about it.

Get in TouchLinkedIn