UniSocial-Secure: High-Performance Social Platform with Custom Engine

📱 Project Overview

UniSocial-Secure is a research-oriented social networking application for Android. Unlike typical apps that rely on SQLite or Room, this project features a custom-built in-memory database engine based on AVL Trees and a zero-dependency CSV persistence layer.

It demonstrates advanced software engineering capability by implementing core data structures and complex design patterns from scratch, optimizing for O(log N) performance and 500,000+ OPS.

Technical Architecture

1. Custom Storage Engine (The Core)

Instead of standard databases, I implemented a high-performance in-memory structure:

• Immutable AVL Tree: Manually implemented self-balancing binary search trees (AVLTree<T>) using structural sharing to guarantee thread safety without explicit locking.
• Performance: Optimized algorithms support data retrieval and insertion at O(log n) complexity.
• Data Slicing: Designed AVLSlice algorithm to support efficient pagination and infinite scrolling in the UI.

2. Content Censorship System (Refactoring & Patterns)

Refactored a legacy monolithic module into a flexible system using the Strategy Pattern:

• Strategies Implemented: BlockProfanity, CaseSensitive, and LeetSpeak (variant detection).
• Dynamic Configuration: Strategies can be swapped at runtime without altering core logic.
• Quality Assurance: Refactoring was validated via rigorous Regression Testing.

3. Architecture & Persistence

• State Pattern: Managed user authentication (Guest/Member/Admin) to eliminate complex conditional logic.
• Data Pipeline: A generic pipeline using the Abstract Factory Pattern to decouple Serialization, Formatting (CSV), and I/O.
• Zero-Dependency CSV Parser: Hand-written Finite State Machine (FSM) parser compliant with RFC standards.

Tech Stack & Skills

• Language: Java 17, Android SDK (Material Design 3)
• Android: RecyclerView, ViewHolders, Intents, ConstraintLayout.
• Engineering: CI/CD (GitLab CI), Unit Testing (JUnit 4 + Hamcrest), Branch Coverage (100%).
• CS Concepts: Recursion, Memory Management, Complexity Analysis.

Key Features

• Secure Persistence: Custom DataPipeline serializes objects to encrypted-style CSV.
• Smart Feed: Optimized feed loading using SortedDataSlice for memory efficiency.
• Live Moderation: Instant filtering of inappropriate content during post submission.

---

Developed as a capstone engineering project to explore manual data structure implementation in mobile environments.