Full Stack Developer
Achievement-driven person with eager to contribute to success through challenging tasks, hard work, and dedication, passionate about learning new technologies and solutions._
Hand-coded since 2021, before ChatGPT came out.
Hi! I'm Egzon Krasniqi, a passionate Full Stack Web Developer from Kosovo. I specialize in building large-scale, robust, and scalable backend systems, with over 4.7+ years of hands-on experience specifically in iGaming solutions, focusing on building sportsbook and wallet systems.
My tech stack focuses on .NET for backend development, combined with Apache Kafka, Microsoft Orleans, and RabbitMQ for event-driven architectures. I am skilled in working with SQL databases and gRPC for high-performance interservice communication. To ensure system reliability and performance, I consistently utilize monitoring tools like Dynatrace, OpenTelemetry, Grafana, and Kibana.
When I'm not coding, you can find me immersed in Motorsports (especially Formula 1), exploring the financial markets, or out on a hike. 🙂
PillHour is a medication and health tracker for iOS and Android, built with Expo and React Native. It works fully offline and without an account: medicines, doses, readings and reminders live in SQLite on the phone, and syncing to a self-hostable .NET backend is something you can turn on later. Adding a medicine is a nine-step wizard that asks one question at a time, from the form and strength to how often and when you take it. The Today screen lists each dose with Take, Not sure, Skip and Later, and every action can be undone. Reminders cover the next 60 doses over a rolling three-week window, Android notifications have Taken, Not sure and Snooze buttons, and an optional persistent alarm uses AlarmKit on iOS 26 and exact alarms on Android through a custom Expo module. Stock is tracked as a ledger with packages and expiry dates, so the app warns you before a medicine runs out. There is also a progress screen with streaks, trend charts for blood pressure, glucose and weight, a doctor report for the last 7, 30 or 90 days, a visit-preparation screen that suggests questions from your own records, and an interaction checker that searches official FDA drug labels once you confirm. Sync goes through an outbox on the device. The server applies each batch in one transaction, keeps receipts so a retry never applies twice, and orders changes by its own sequence number, never by device clocks. When two devices disagree, the server's version wins and the local copy is saved in a conflicts table on the phone. The app is translated into 27 languages, including Albanian and the right-to-left Arabic and Urdu.
Ura means "bridge" in Albanian. It is a job board and hiring tool built for Kosovo, in Albanian first and English second. Job seekers search jobs across all 38 municipalities, see the pay wherever the employer publishes it, apply with one profile and CV, and follow each application through its stages. Employers register their company, post jobs, and work through applicants as a team. Search runs inside PostgreSQL with a full-text index and trigram matching, and it understands how people actually type: "prishtine" finds "Prishtinë", and a sentence like "gjysmë orari në Prizren mbi 500 euro" (part-time in Prizren, over 500 euros) turns into the matching filters. Seekers also get a CV builder with three templates in Albanian, English or German, a panel that compares their profile with each ad, saved searches, email alerts, and a salary calculator that converts gross to net pay under Kosovo's tax and pension rules and shows the legal source for each line. Employers get a dashboard and an applicant pipeline with stages, ratings, private notes and mentions, shared with the recruiters they invite by email. Ads from unverified companies go to a moderation queue, where rule-based checks flag things like asking candidates to pay. Sign-in is passwordless with a six-digit code sent by email. The backend is .NET 10 with EF Core, PostgreSQL, Redis and a worker for emails, alerts and data retention, and the frontend is Next.js 16 with React 19, Tailwind and next-intl.
Plarchy is a house-plan app for people planning a home of their own. You list the rooms you need, give the house a rough shape and a lot size, and Plarchy turns that brief into five floor plans with real walls, doors and windows. Pick the one you like and it furnishes the plan, renders the exterior with AI, and gives you drawings a builder or architect can open. Every plan is a built model. Walls sit on shared centerlines, rooms are measured to the sixteenth of an inch, and a layout planner follows the rules a residential designer would: bedrooms off a hallway, the kitchen open to the dining room, the garage on a corner with a door into the house. The floor plan, the 3D model, the AI renders and every export come from that one model, so they always describe the same house. Each design page reads the plan back to you. It shows a room schedule, whether the house fits its brief and its lot, how much glass each bedroom gets, how far you walk from the garage to the kitchen, which doors are too narrow for a wheelchair, and a takeoff of walls, roof, floors, doors and windows. You can download the plan as a PDF plan set, a DXF for AutoCAD, Revit or SketchUp, a GLB model for Blender, or an IFC file for BIM tools. The backend is .NET 10 with PostgreSQL, Redis and a worker that runs the generation jobs, and the frontend is Next.js 16 with React Three Fiber for the 3D views.
PropFormIQ is a web app for modelling real estate deals. An analyst enters a project's costs, income, timing and financing, and the app works out whether the deal makes money: NPV, IRR, equity multiple, profit on cost, loan coverage, and the cash flow year by year. It is live at propformiq.com with a free trial and paid Professional and Team plans. There are seven property types, each with its own model: retail, office, hotels, homes for sale, rental property, mixed use, and entertainment venues. Every result traces back to a named input, and each input says in plain words what it does. From there you can test assumptions with one-way and two-way sensitivity tables, save alternatives, compare versions, check the project against benchmarks and published market data, and roll several projects up into a portfolio. One saved calculation feeds every output: a PDF investment memo you put together section by section, an Excel workbook, a PowerPoint committee deck and a CSV. A project assistant answers questions about the results and cites every number it uses, but it never calculates anything itself. The calculation engine is plain C# on .NET 10 and has about 2,700 back-end tests. The frontend is Next.js 16, sign-in runs through Keycloak, billing through Stripe, and the data lives in PostgreSQL. It runs in Docker on a VPS, deploys from CI on every push to main, and the interface is translated into 28 languages.
KosHome is a specialized real estate marketplace designed specifically for the Kosovo market, bridging the gap between property seekers and landlords. The platform features an interactive map-based search where listings appear as dynamic pins, allowing users to filter properties by city, price, and amenities in real-time. For buyers and renters, it offers detailed listing pages with image galleries, location analytics, and integrated mortgage calculators. For property owners and agents, it provides a comprehensive dashboard to manage listings, track engagement metrics like views and specific inquiries, and leverage AI tools for generating descriptions and estimating market values. Built with a clean architecture on .NET 9, it uses PostgreSQL for data handling and Keycloak for secure, role-based authentication.
A microservices platform that reads financial news from across the web, figures out whether the sentiment is positive or negative using AI, and compares that to how cryptocurrency prices are actually moving. When big news drops, the market reacts fast. This system catches that early and sends out alerts so traders can stay ahead. I broke the project into four separate apps that each handle one job, like fetching news, analyzing it, or matching it to price data. They all talk to each other through message queues, which means if one part gets busy, it does not slow down everything else. The frontend is a dashboard built with React and Next.js where users can track everything in real-time.
I built this Ticketing platform to explore how modern system architecture can prevent the common problem of websites crashing during high-demand events. Instead of a single, bulky application, I used a microservices approach where different sections like payments, orders, and event listings operate independently. This ensure that a surge of traffic in one area doesn't bring the whole site down. By using Docker and Kubernetes, I made the entire infrastructure elastic, allowing it to scale resources up and down automatically based on real-time visitor numbers. It is a highly resilient system designed to handle the intense pressure of a global ticket marketplace, proving that thoughtful engineering in the background can create a consistently fast and unbreakable experience for every user in various scenarios.
Activity Wave is a social platform I created to solve the problem of finding interesting things to do in your local area. I wanted to move away from the usual passive scrolling and build a space that actually encourages people to get out and connect in person. The app allows anyone to host an activity from a weekend football match to a casual study group while others can browse and join nearby events. To make things easier, I included a real-time chat for every activity so participants can coordinate before meeting up. It also features a social layer where you can follow friends, customize your profile, and see what events your network is joining, making the whole experience feel like a lively, active community rather than just a simple event listing site.
With Crown Clothing, I set out to build a platform that users could truly trust. This project was a deep dive into creating a professional-grade authentication and state management system that stays solid even as the store grows. Using React and Redux, I made sure that managing a dynamic cart across several categories never slows down the experience. One of the highlights is the advanced security layer, featuring a robust JWT-based authentication system with both access and refresh tokens a standard usually found only on large global platforms. Whether visitors are signing in through Google or managing their personal accounts, they can shop with confidence knowing their data is safe. It is a premium, boutique-style destination where high-end technology meets a beautiful and intuitive fashion experience.
When I built Clothesy, I wanted to show that online shopping for clothes should be as light and simple as possible. Modern fashion sites are often cluttered, so I took a minimalist approach to keep the focus entirely on the apparel. Using Vue and Vuetify, I created an interface that feels especially smooth on mobile devices, allowing users to browse curated collections and view high-quality images without any lag. By using Firebase for the backend, I made sure that every interaction from adding an item to the cart to checking the latest arrivals stays updated in real-time. I also integrated Stripe to handle secure payments, ensuring that the journey from finding a piece of clothing to finalizing the purchase is frictionless, fast, and completely safe for every customer.
Managing a cinema involves managing everything from movie schedules to snack bar inventory, so I built CinemaBooking to handle all those moving parts in one place. It is a desktop management system that makes it easy for theater staff to run their daily operations efficiently. For moviegoers, I focused on creating a smooth booking journey where they can see what is currently playing, pick their specific seats from a visual map, and even pre-order snacks to avoid the lines. On the administrative side, the app includes a powerful dashboard for tracking sales hits and managing different screening halls. I chose Java and JavaFX for this project because I wanted to show that professional desktop tools can be both powerful and visually appealing, making the whole movie-going experience more organized for both the staff and the customers.
Clothbox is a high-performance online store I built to explore the power of modern web technologies like GraphQL. My goal was to create a shopping experience that feels incredibly fast and responsive, whether you are browsing sneakers or managing a massive inventory. I integrated secure sign-ins with Google to keep the entry barrier low, and used Stripe to ensure that every payment is safe and reliable. Beyond the storefront, I built a detailed admin dashboard where owners can easily update products, track orders in real-time, and organize the catalog without touching any code. It is a complete e-commerce solution that combines a clean, premium fashion aesthetic with a scalable technical foundation, proving that professional digital stores can be both beautiful and highly functional.
I built JobGuide to tackle the stress of the career search by creating a reliable bridge between talented people and the companies that need them. For job seekers, I focused on making the hunting process feel intuitive and organized, allowing them to filter through listings and submit applications in just a few clicks without unnecessary distractions. For employers and administrators, I developed a powerful suite of management tools to post new openings and track candidates through the entire hiring pipeline in a organized way. By leveraging a robust .NET Core backend, I ensured that the entire platform stays fast and stable even as more users sign up. It is a professional hub designed to turn a complicated and often frustrating task like finding a new role into a very smooth and straightforward journey for everyone involved.
A human guide to building a scalable deployment pipeline using Kubernetes, Kustomize for environment management, and ArgoCD for GitOps.
Learn how to implement Keycloak authentication in .NET 10 using the Options pattern for a clean and scalable security configuration.
A practical deep dive into what the C# compiler generates for async/await, and why it matters in real code.