JK
Projects
JK
>LOADING // LOADING
INITIALIZING VIEW
Built and Designed byJK
JK
>LOADING // PROJECT DETAILS
PARSING ARCHITECTURE, DIAGRAMS & API ENDPOINTS
Back to Projects
Live DemoView on GitHub
LaTeX Resume Generator
Completedreacttypescriptvitetailwindgogindocker

LaTeX Resume Generator

A frontend-first, privacy-focused resume builder for IT professionals that generates ATS-friendly, single-page A4 LaTeX resumes. The frontend handles all data editing, live preview, and template rendering entirely in the browser with localStorage persistence. A lightweight Go backend exists solely to compile LaTeX to PDF via Tectonic.

Problem & Solution

Building publication-quality resumes typically requires LaTeX knowledge. This project solves that by letting users fill structured forms, choose from 9 ATS-optimized templates, see a live HTML/CSS preview, and export to either .tex source or compiled PDF — all without writing a single line of LaTeX.

Features

  • 9 pluggable LaTeX templates with live HTML/CSS preview
  • Frontend-first architecture with localStorage persistence
  • Real-time WebSocket PDF compilation with progress streaming
  • Drag-and-drop section reordering via @dnd-kit
  • Profile image upload with cropping (react-easy-crop)
  • Cmd+K command palette for power users
  • Version management with save/load/export
  • Keyboard shortcuts (Ctrl+P, Ctrl+L, Ctrl+S, Ctrl+D)
  • Dark/light mode toggle
  • Multi-page preview simulation with overflow detection
  • Sentry error tracking with session replay
  • Docker Compose with Prometheus + Grafana observability

Tech Stack

TechnologyVersionPurposeWhy Chosen
React19UI frameworkLatest concurrent features, lazy loading
TypeScript~6.0Type safetyPrevents runtime errors on complex resume data model
Vite8.xBuild tool & dev serverFast HMR, native ESM, code splitting via glob imports
Tailwind CSSv4Utility-first stylingRapid UI development, consistent design tokens
Zustand5.xState managementMinimal boilerplate, built-in persist middleware
@dnd-kitcore 6, sortable 10Drag-and-dropAccessible, performant sortable lists
Go1.25+Backend runtimeFast startup, small binaries, excellent concurrency
Gin1.12HTTP frameworkHigh performance, middleware ecosystem
Tectonic0.16.9LaTeX→PDF compilerSelf-contained, no TeX distribution needed
Docker—ContainerizationReproducible builds, production parity
SQLitemodernc.org/sqlite 1.54Stats persistenceNo CGO dependency, cross-compilation
Sentry10.xError tracking & monitoringSession replay, performance monitoring, feedback widget

Architecture

React 19 SPA

Zustand Store

localStorage

Live HTML Preview

LaTeX Generator

9 Template Engine

Gin Server

:8080

Rate Limiter

API Key Auth

Compile Handler

Tectonic

SQLite Stats

Prometheus

Grafana

Press enter or space to select a node. You can then use the arrow keys to move the node around. Press delete to remove it and escape to cancel.
Press enter or space to select an edge. You can then press delete to remove it or escape to cancel.

Frontend (React SPA)

  • React 19 SPA with Zustand state management
  • 9 pluggable template strategy pattern (config + preview + LaTeX generator)
  • Live A4 page simulation with zoom and pagination
  • WebSocket-based real-time compilation streaming
  • Vite 8 with code splitting via import.meta.glob

Backend (Go Microservice)

  • Go/Gin HTTP server with single concern: LaTeX compilation
  • Tectonic subprocess management with --untrusted sandboxing
  • Per-IP token bucket rate limiting (5 rps, burst 10)
  • API key authentication on compile endpoints
  • SQLite for aggregate visit/download statistics

Infrastructure

  • Docker Compose orchestration (backend + Prometheus + Grafana)
  • Multi-stage Docker builds (golang → alpine + tectonic)
  • Vercel deployment for frontend, Render for backend
  • Prometheus metrics with Grafana dashboards

Project Structure

backend
cmd
server
main.goGin server setup, routes, middleware
internal
compiler
handlers
middleware
websocket
frontend
src
components
editor
preview
hooks
stores
templates

API Endpoints

Go Backend (port 8080)

MethodRoutePurpose
POST/api/compileCompile LaTeX to PDF (HTTP)
GET/api/compile/wsWebSocket compile with progress streaming
POST/api/stats/visitRecord page visit
POST/api/stats/downloadRecord PDF download
GET/api/statsGet aggregate stats
GET/api/healthHealth check (Tectonic + DB status)
GET/metricsPrometheus metrics

Database Design

stats

2 cols
keyPK
TEXT
value
INTEGER
Press enter or space to select a node. You can then use the arrow keys to move the node around. Press delete to remove it and escape to cancel.
Press enter or space to select an edge. You can then press delete to remove it or escape to cancel.

stats

ColumnTypeNotes
keyPKTEXTCounter name: 'visits' or 'downloads'
valueINTEGERCounter value, default 0

Single table with 2 key-value rows. WAL mode enabled. Used only for aggregate visit/download statistics — no user data stored.

No relationships. Single table with 2 counter rows.

Key Design Decisions

Frontend-First Architecture

Decision: All resume data, editing, preview, and LaTeX generation happen in the browser

Why: Privacy-first: all data stays in browser with localStorage persistence

Benefits: Zero server costs for editing, user privacy, offline-capable editing

Trade-offs: No cloud sync, no cross-device access, localStorage size limits

Tectonic for LaTeX Compilation

Decision: Use Tectonic instead of pdflatex or other TeX distributions

Why: Self-contained binary, no TeX distribution needed on the server

Benefits: Simple Docker image, reproducible builds

Trade-offs: Slower than pdflatex for complex documents

Zustand over Redux/Context

Decision: Use Zustand for state management with built-in persist middleware

Why: Minimal boilerplate, built-in localStorage persistence

Benefits: Simple state management, easy persistence

Trade-offs: Less ecosystem than Redux

WebSocket for Compile Streaming

Decision: Use WebSocket instead of HTTP polling for PDF compilation

Why: Real-time progress feedback improves UX during 5-60s compilation

Benefits: User sees compilation stages (validating → writing → compiling → done)

Trade-offs: More complex than HTTP polling

Strategy Pattern for Templates

Decision: Each template implements a Template interface with config, Preview, and generateLatex

Why: Multiple visual styles with shared data model

Benefits: Extensible, each template self-contained, lazy-loadable via import.meta.glob

Trade-offs: Code duplication across templates

SQLite for Stats Only

Decision: Use SQLite for aggregate visit/download counters only

Why: Zero-config embedded DB adequate for 2 counter rows

Benefits: No external database dependency, simple deployment

Trade-offs: Write contention under concurrent access

Testing

None configured•0 unit tests

What's Tested

    Missing Coverage

    • Unit tests for escapeLatex, generateLatex functions
    • Unit tests for template rendering and LaTeX generation
    • Integration tests for compile handler
    • Integration tests for WebSocket compilation flow
    • E2E tests for critical user flows (edit → preview → export)

    Key Challenges

    • LaTeX compilation timeout (Tectonic can take 60+ seconds for complex documents)
    • Profile image size in request body increasing payload and compilation time
    • localStorage quota limits (~5MB) for saved resume versions
    • Multi-page PDF warning — preview is approximate, actual PDF may differ
    • Concurrent compilation on single-container deployment

    Key Learnings

    • Strategy pattern for pluggable template architecture
    • WebSocket streaming for real-time compilation progress
    • LaTeX escaping with null-byte placeholders to prevent double-escaping
    • Multi-page preview simulation with section-aware pagination
    • Frontend-first architecture trade-offs (privacy vs. no cloud sync)