Project Portfolio

Developed a secure, transaction-safe backend API that orchestrates user authentication, role-based permissions (for customers, providers, and admins), and service booking workflows. It resolves critical backend challenges around stateless session security (JWT + logout blacklist), database schema integrity (via versioned migrations), and API integration reliability (via Swagger UI documentation and extensive JUnit testing).

Built a full-stack movie and TV streaming platform that enabled users to search, browse, and manage entertainment content efficiently while providing a fast, mobile-friendly, and SEO-optimized user experience.

The project solves the problem of automatically identifying the species of an Iris flower based on its physical measurements. Traditionally, flower classification requires botanical knowledge and manual inspection. This machine learning solution analyzes sepal and petal dimensions and predicts whether the flower belongs to the Setosa, Versicolor, or Virginica species. The project demonstrates the complete machine learning workflow, from data preparation and model training to evaluation and deployment as an interactive web application, making flower classification fast, accurate, and accessible to users without expertise in botany.

Developed, maintained, and optimized a production Laravel web application by diagnosing and resolving critical issues affecting performance, user experience, authentication, search functionality, SEO visibility, and system stability. Improved application reliability through debugging, performance tuning, server-side rendering enhancements, database query optimization, and deployment troubleshooting.

Students at Addis Ababa University often struggle to find affordable, trustworthy help for academic, creative, and daily life tasks. Existing options rely on scattered WhatsApp groups, word of mouth, or informal networks, which lack verification, structure, and safety. Many students possess valuable skills such as tutoring, editing, design, or delivery but have no centralized, trusted environment to offer their services or build a portfolio. At the same time, students who need help cannot easily discover reliable peers, compare options, or ensure they are dealing with verified individuals. Financially, many students face limited allowances and lack flexible ways to earn small but meaningful income on campus, despite having marketable skills. This creates a gap where students who could support each other academically and economically remain disconnected. A dedicated, student‑only system with identity verification and campus‑based organization is needed to address these academic, social, and financial challenges within the AAU community.

Creating highly detailed, 3D geopolitical animations using standard video editing software is incredibly slow and lacks real geographic accuracy. I solved this by engineering a programmatic map animation engine that layers custom 3D tactical graphics (Three.js) over a mathematically accurate 3D globe (CesiumJS), allowing for the rapid generation of cinema-grade, geographically precise military and geopolitical simulations.

Traditional motion graphics workflows using software like After Effects are highly manual, time-consuming, and difficult to scale. I engineered a programmatic Motion Graphics Creation System using React and Remotion to solve the bottleneck of manual video production. This system allows for the automated generation of cinematic, data-driven animations, such as dynamic typography sequences, complex text morphing, and high-volume cinematic credits. By defining animations through code, the system ensures pixel-perfect brand consistency, enables instant updates to text or data without manual re-rendering, and dramatically scales up production speed for broadcast-ready visual assets.

Tested the AiKiTech LLM Router platform to verify intelligent routing between multiple Large Language Models using developer-provided API keys. The project involved functional API validation, routing analysis, concurrent load testing, performance measurement, latency analysis, cost comparison, and developer experience evaluation.

BETA TESTING Aliens vs Zombies on Android App to better understand first session of players to understand what the player does not understand and what is the logic of the behavior of new players. THE SPECIFIC APP VERSION TO BE TESTED IS: App version in google play is 2411 12 126 SO Evaluated the first-time player experience of an Android mobile game by identifying usability issues, unclear game mechanics, onboarding friction, and player behavior during the initial gameplay session. The objective was to help the development team understand how new players interacted with the game and where confusion or drop-off occurred.

Evaluated the usability, functionality, and first-time user experience of a digital currency trading platform. The objective was to assess how easily new users could create digital currencies, trade with other users, manage wallets, and navigate the platform while identifying usability improvements and potential user adoption barriers.