LEDO-TECH — Enterprise Software Development Company

Universal File Converter Pro — Engineering Case Study

Flagship Desktop Application | Built with C# and .NET 8

System Architecture Overview

Universal File Converter Pro is a native Windows desktop application built on the .NET 8 runtime, designed to handle batch file conversion across more than 50 file formats — including video, audio, image, and document types. The application leverages a modular pipeline architecture where each conversion task is decomposed into discrete stages: input validation, codec detection, format-specific decoding, transformation, encoding, and output verification.

Multithreaded Processing Engine

At the core of the application is a custom-built Task Parallel Library (TPL) pipeline that distributes conversion workloads across all available CPU cores. Each file in a batch job is processed in its own thread-safe context using Task.Run() with configurable concurrency limits via SemaphoreSlim. This architecture prevents memory exhaustion during large batch operations (500+ files) while maintaining near-linear throughput scaling on multi-core processors.

The stream processing subsystem uses memory-mapped file I/O (MemoryMappedFile) for large files exceeding 100MB, avoiding the overhead of loading entire files into heap memory. For smaller files, buffered FileStream operations with 64KB read/write buffers provide optimal throughput. The engine dynamically selects the appropriate I/O strategy based on file size heuristics determined during the input validation stage.

Format Matrix & Codec Management

The format compatibility matrix supports over 50 input/output format combinations, including MP4, AVI, MKV, MP3, FLAC, WAV, PNG, JPG, WebP, AVIF, PDF, DOCX, and more. Each format is handled by a dedicated codec adapter implementing a common IFormatConverter interface, enabling clean separation of concerns and easy extensibility. Video and audio codecs leverage FFmpeg interop through a managed wrapper, while image conversions use System.Drawing and SkiaSharp for hardware-accelerated pixel manipulation.

Privacy & Security Architecture

A critical design principle of Universal File Converter Pro is its zero-data-upload privacy model. All file processing occurs entirely on the user's local machine — no files are transmitted to external servers. This architecture not only eliminates privacy concerns but also removes network latency from the conversion pipeline, enabling offline operation and significantly faster throughput compared to cloud-based converters.

Learn more about Universal File Converter Pro →

Ultra Resizer — Browser-Based Image Processing Suite

Web Application | HTML5 Canvas API, JavaScript, Client-Side Processing

Canvas API & Image Manipulation Pipeline

Ultra Resizer is a professional-grade image processing toolkit that runs entirely in the browser using the HTML5 Canvas API and modern JavaScript APIs. When a user uploads an image, it is decoded into a raw pixel buffer on an offscreen CanvasRenderingContext2D context. All resize, crop, rotate, and filter operations are performed as direct pixel manipulations on this buffer — using bilinear interpolation for upscaling and Lanczos resampling for downscaling to maintain visual fidelity.

4K Upscaling & Smart Compression

The 4K upscaling feature uses multi-pass sharpening algorithms applied after initial bilinear interpolation. Each pass applies a convolution kernel that enhances edge definition while suppressing noise artifacts that typically appear when scaling images beyond their native resolution. The compression engine supports output in JPEG (with configurable quality from 1-100), PNG (lossless), WebP (both lossy and lossless modes), and AVIF — with real-time file size previews before download.

AI Background Removal

The background removal tool uses a client-side machine learning model to segment foreground subjects from their backgrounds. The model runs inference directly in the browser using WebAssembly, processing each image through a neural network that generates a pixel-level segmentation mask. This mask is applied to the original image to produce a transparent-background PNG output — all without sending user images to any external server.

Privacy Guarantee

Ultra Resizer processes all images using client-side JavaScript and the HTML5 Canvas API. No images are uploaded to any server. All processing occurs locally in the user's browser, ensuring complete privacy and data sovereignty. The application works offline after initial page load and imposes no limits on the number of images processed.

Try Ultra Resizer for free →

Pulsgram — Mobile Step Counter & Gamification Platform

Mobile App | .NET MAUI, Supabase, Real-Time Database

Cross-Platform Architecture with .NET MAUI

Pulsgram is a cross-platform mobile application built with .NET MAUI (Multi-platform App UI), enabling a single C# codebase to target both Android and iOS. The app's primary function is step counting with gamification — users walk to grow a virtual tree through 10 unique growth stages, compete on a global leaderboard, and challenge friends to 1v1 step battles. The architecture follows the MVVM (Model-View-ViewModel) pattern with dependency injection for service resolution.

Step Detection & Sensor Fusion

The step counting algorithm uses the device's hardware accelerometer sensor, processed through a custom peak-detection filter that identifies the rhythmic gait pattern characteristic of walking, running, and climbing. To optimize battery consumption, the app uses an adaptive polling strategy: high-frequency sampling (10Hz) during active tracking sessions and low-frequency polling (0.5Hz) during background operation. The background service is implemented as an Android Foreground Service with a persistent notification to prevent OS-level process termination.

Supabase Integration & Real-Time State

Pulsgram uses Supabase as its backend infrastructure, leveraging PostgreSQL for persistent data storage and Supabase Realtime for live leaderboard updates. When a user's step count changes, the update is pushed to Supabase via a debounced API call (batched every 30 seconds to minimize network overhead). The leaderboard subscription uses WebSocket channels, enabling instant rank updates across all connected clients without polling.

Virtual Tree Growth & Gamification Engine

The gamification system maps cumulative step milestones to tree growth stages (Seedling → Sapling → Oak → Ancient Tree). Each growth stage unlocks new visual assets and achievement badges. The progression curve follows a logarithmic function to ensure early gratification while maintaining long-term engagement. The 1v1 step battle system uses Supabase's Row Level Security (RLS) to ensure fair play and prevent fraudulent step injection.

Explore Pulsgram →

Orbital & Observatory Tracker Pro — 3D Celestial Simulation

Interactive 3D Web App | Three.js, WebGL, SGP4 Orbital Mechanics

Three.js WebGL Rendering Pipeline

Orbital & Observatory Tracker Pro renders a full 3D model of the solar system in the browser using Three.js and WebGL. The scene graph includes accurately scaled planetary bodies with physically-based rendering (PBR) materials, orbital paths computed from Keplerian elements, atmospheric scattering shaders, and a high-resolution star field backdrop generated from the Hipparcos star catalog. The rendering pipeline targets 60fps with automatic quality adjustment based on GPU performance metrics.

SGP4 Satellite Propagation

Real-time satellite tracking uses the SGP4 (Simplified General Perturbations 4) orbital propagation model, the same algorithm used by NORAD and NASA for tracking objects in Earth orbit. The application ingests Two-Line Element (TLE) sets from CelesTrak's public API and computes satellite positions in Earth-Centered Inertial (ECI) coordinates at each animation frame. These positions are transformed to Earth-Centered Earth-Fixed (ECEF) coordinates and then to Three.js world space for accurate 3D rendering above the Earth model.

ISS Live Tracking & Telemetry

A dedicated ISS tracking module provides real-time position data for the International Space Station, including latitude, longitude, altitude, velocity, and orbital period. The ISS position is updated every second using SGP4 propagation from the latest TLE data, with visual indicators showing the current footprint (visible area from the station) and ground track projection on a 2D Mercator map overlay.

Celestial Ephemeris System

The ephemeris engine computes planetary positions using VSOP87 theory for the inner planets and simplified Keplerian solutions for outer planets. This enables accurate visualization of planetary alignments, conjunctions, and oppositions across user-selectable time ranges. The time control system allows users to scrub forward and backward through orbital time at adjustable speeds (1x to 100,000x real-time).

Launch Orbital Tracker Pro →

Universal Calculator — Multi-Mode Calculation Suite

Web Application | JavaScript, Responsive Design

Multi-Mode Calculation Engine

Universal Calculator provides a comprehensive calculation suite with multiple operating modes: Standard (basic arithmetic with order of operations), Scientific (trigonometric, logarithmic, exponential, and factorial functions), and Programmer (binary, octal, hexadecimal conversions with bitwise operations). Each mode shares a common expression parser that tokenizes input, builds an abstract syntax tree (AST), and evaluates results with floating-point precision handling to minimize rounding errors.

Built-in Unit Converter

The integrated unit converter supports conversions across 12 categories — including length, weight, temperature, speed, area, volume, time, energy, data storage, pressure, angle, and force. Each category uses a base-unit normalization approach where all input values are first converted to a canonical base unit (e.g., meters for length, kilograms for weight) and then converted to the target unit, ensuring conversion accuracy across all unit pairs without maintaining an explicit conversion matrix.

History & Memory System

The calculator maintains a scrollable history of all computed expressions and results, persisted in browser localStorage for cross-session continuity. The memory system supports standard M+, M-, MR, and MC operations with support for multiple memory registers, enabling complex multi-step calculations.

Open Universal Calculator →

AI Resume Pro — AI-Powered Resume & CV Builder

Web Application | JavaScript, AI Integration, PDF Generation

AI-Powered Content Generation

AI Resume Pro integrates with large language model APIs via OpenRouter to provide three AI-powered features: automatic professional summary generation based on user-provided job titles and experience, intelligent experience description improvement that enhances bullet points with action verbs and quantified achievements, and context-aware skill suggestions that recommend relevant skills based on the user's industry and role. The AI system prompt is carefully constrained to resume and CV domains only, preventing off-topic generation.

Template Engine & Real-Time Preview

The application ships with 7 professionally designed resume templates — Modern, Classic, Executive, Technical, Academic, Minimalist, and Creative — each implementing a distinct visual layout while sharing a common data model. Templates are rendered in real-time as the user types, using a reactive state management system that re-renders the A4-proportioned preview document on every keystroke. Users can switch templates, theme colors, typography, and background effects without losing any entered data.

PDF Export Pipeline

The PDF export feature uses html2pdf.js — a client-side JavaScript library that captures the rendered HTML resume document, converts it to a high-resolution canvas, and generates a multi-page PDF using jsPDF. The export pipeline preserves all CSS styling, custom fonts, profile images, and layout proportions at 300 DPI equivalent resolution. All PDF generation occurs in the browser — no server involvement, ensuring user resume content remains completely private.

ATS Optimization

All resume templates generate clean, semantic HTML structure that translates to ATS-friendly (Applicant Tracking System) PDF output. The Minimalist template in particular is designed to maximize ATS compatibility by using a single-column layout, standard section headings, and avoiding complex formatting elements that ATS parsers commonly fail to interpret.

Build your resume with AI →