WordNoteWordNote

Backend Developer Portfolio

Danh Son Ha

I build backend systems that turn product ideas into reliable APIs, data models, review engines, and deployed services. My strongest work is WordNote, an AI-powered vocabulary platform with spaced repetition, OAuth, quiz validation, and adaptive review logic.

Contact meGitHubView WordNote
Go
backend services
PostgreSQL
data modeling
OAuth
auth flows
SRS
learning engine

Backend signal

Production-minded engineering, not just feature wiring.

I care about API contracts, data correctness, fallback behavior, and testable domain logic. The projects below show work across authentication, queues, review scoring, AI validation, deployments, and product-facing backend decisions.

Featured project

WordNote - AI vocabulary learning platform

A full-stack English learning product where users capture real vocabulary from web context, receive AI explanations, and review words through spaced repetition.

Backend ownershipwordnote.id.vn

Built the core backend for capture, AI enrichment, review sessions, and user learning stats.

  • Go REST API with clean usecase, repository, and domain boundaries.
  • PostgreSQL schema migrations for words, reviews, sessions, quiz validation, and user learning stats.
  • JWT and Google OAuth flows for web and Chrome extension clients.
  • AI vocabulary pipeline that generates explanations, examples, quiz variants, and validation-safe review content.
  • Spaced-repetition queue, usable scoring, mistake recovery, and adaptive session planning.
  • Table-driven tests around scoring, quiz validation, session behavior, and sparse-data edge cases.

Architecture

How WordNote turns saved words into reviewable learning data

Capture

Context-first vocabulary ingestion

Words are saved with the sentence and source where the user found them, then normalized into structured learning data.

Enrich

AI pipeline with validation

The backend turns raw words into definitions, examples, CEFR metadata, quizzes, and rejectable invalid content.

Review

Adaptive SRS sessions

Review queues prioritize due words, mistakes, low usable score, production gaps, and balanced quiz difficulty.

Measure

User-word learning signals

The system tracks review outcomes, quiz coverage, time gap, production accuracy, and memory score as usable ability.

Experience

Advanced Computing Lab

IoT backend, July 2024 - July 2025

Designed API contracts and message schemas for reliable data exchange between edge devices and cloud services.

Built device-to-cloud communication services over MQTT and HTTP for real-time telemetry ingestion.

Implemented connectivity flows across WiFi, BLE, and cloud integrations for IoT systems.

Stack

Technologies I can defend in an interview

Backend

GoGinnet/httpREST APIClean Architecture

Data

PostgreSQLSQL migrationsMongoDBRepository Pattern

Auth & Security

JWTGoogle OAuthCORSrate-limit planning

Production

RailwayVercelDockerCI/CDstructured testing

AI Systems

OpenAIGeminiDeepSeekprompt validationcontent quality gates

Contact

Looking for backend roles where product logic matters.

Ho Chi Minh City, Vietnam. Open to backend engineering roles focused on APIs, data modeling, production systems, and learning or AI products.

sonha2002.12@gmail.com+84 357 607 619