AJ Tran

Full-stack software engineer. Enterprise scale by day. Personal scale by night.

View Projects
AJ Tran

About Me

Hey, I'm AJ!

By day: Ruby, Python, Java, TypeScript. I build systems that scale, modernize code that doesn't, and maintain the kind of test coverage that lets you sleep at night.

By night: NLP on album lyrics, computer vision for dance analysis, CLI tools for vinyl collectors. Projects that start with "What if?" and end with working code.

Featured Projects

CyberStudy Scheduler

Summary

A personal study tracker built for cybersecurity coursework with spaced repetition scheduling and AI agent integration. Tracks courses, topics, assignments, and exams through a four-tier mastery system (NOT_STARTED through MASTERED) driven by SM-2 spaced repetition. An autonomous AI agent ("Hex") connects via REST API to run nightly quizzes, Teach It Back sessions, and daily study pings — creating a closed-loop system where wrong answers compound through two independent feedback loops.

View the code on GitHub

Highlights

  • SM-2 Spaced Repetition: Topics at LEARNING or PROFICIENT participate in review scheduling. Each review takes a quality rating (0–5) and computes the next review date.
  • Daily Briefing Engine: Surfaces what's due, what's coming up, and what to focus on — the single entry point for "what should I study now?"
  • AI Agent Integration: Full REST API designed for an autonomous study agent that generates quizzes, grades teach-back sessions, and syncs key terms from Obsidian notes.
  • Dual Feedback Loops: Quiz misses weight future questions toward weak topics while SM-2 independently surfaces those same topics at shorter review intervals.
  • Review Queue: Multiple sort modes including mastery priority, SRS due date, interleaved (cross-course round-robin), and least-recently-reviewed.
  • 3D Flashcards: Per-topic key terms with flippable card UI, shuffle, and keyboard navigation.
  • Calendar Views: Week and month views aggregating topics, assignments, and exams in one place.
  • Batch Import: Populate an entire course in one API call with HTTP 207 partial-success support.

Tech Stack

  • Frontend: Next.js 16 (App Router), React 19, TypeScript
  • Database: Prisma 7, PostgreSQL
  • Auth: NextAuth v5 (credentials + JWT), dual-mode Bearer token and session auth
  • UI: shadcn/ui, Tailwind CSS v4
  • Visualization: Recharts
  • API Docs: Swagger UI

Status

In active use for cybersecurity coursework. AI agent runs on a cron schedule handling daily study pings, quiz generation and grading, Teach It Back sessions, and key terms sync from Obsidian.

Daily briefing dashboard showing study priorities and SRS reviews due

1 / 3

Project 1 / 9