TOM
KLUSKENS
About
Full Stack
Developer
I'm a full stack developer finishing my Bachelor's in Applied IT at HoGent, with a thesis exploring AI and energy efficiency. This September, I'll be starting as a software engineer at AE.
To me, engineering is about more than writing code. I care about the full picture: understanding requirements, shaping ideas, and building systems that are solid and easy to maintain. I'm comfortable talking directly with clients, explaining technical concepts in plain language, and making sure the end result actually matches what the business needs.
Production experience, early. I first interned at Turtle Srl in Cesenatico, Italy during my final year of secondary school. They invited me back for my Bachelor's internship — both experiences put me in a real professional environment where I had to adapt quickly and deliver.
Always building. Outside of work, I run my own homelab and personal server. I enjoy experimenting with new technologies and side projects. It keeps my skills sharp and my curiosity alive.
Team player. Years of competitive basketball taught me how to collaborate under pressure, communicate clearly, and support the people around me. I bring that same mindset to every team I work with.
I love building things from the ground up, with a particular interest in frontend work, databases, and infrastructure. I'm always looking for the next opportunity to learn and deliver real value.
Connect on LinkedIn →Projects
9Turns audio recordings of business interviews into structured Word documents. Transcribes MP3s via WhisperX, cleans with an LLM, lets users review and edit, then generates a full report by answering template paragraphs from transcriptions and vector retrieval.
Microservice that takes a search query, retrieves the top 10 results, scrapes and processes page content, indexes it as vectors, and returns semantically relevant results. Repeated queries are served from cache. Includes an LLM agent chat interface for natural-language product search.
Async FastAPI service that discovers pages via sitemaps, filters candidate URLs with deterministic rules + LLM assistance, scrapes with BeautifulSoup/Playwright, and stores results in MongoDB. Rate-limited, API-key protected, deployable on Scaleway serverless or Docker.
Machine status dashboard per site with KPIs, built for Delaware in a team of 5. Full-stack with a Java micro-service layer.
Social media platform for people with mental difficulties. Built in a team of 6 at HoGent — clean architecture with Blazor WASM frontend, ASP.NET 9 API, real-time chat via SignalR, and push notifications.
Conference event management system handling speakers, locations, and user registrations. EWD exam project at HoGent.
A social platform for students to upload, share, like and comment on recipes — with friends and discovery features.
Data visualization dashboard converting customer data into charts. Built during my internship in Cesenatico, Italy.
A school library system that automatically generates QR codes with unique ISBN numbers for every book.
Experience
Developer
Starting as a developer at AE — a Belgian IT consultancy focused on digital transformation, data and software engineering.
Intern · Bachelor Thesis on AI & Energy
Returning to Turtle Srl to complete my bachelor thesis on AI & Energy Efficiency in a live production environment.
Student Worker
Fast-paced customer service and team collaboration. Learned how to stay sharp and efficient in demanding environments.
Intern · Data Visualization
Built a data-visualization dashboard in a live production environment during my first internship abroad.
Student Worker
Customer service and retail operations. Developed reliability, communication and working under pressure.
Skills
Thesis
AI & Energy Consumption in Web Architecture: A Comparison Between Serverless, Classic Infrastructure, and Progressive Web Apps
The rapid growth of AI is driving up global energy consumption. Companies like Turtle Srl need ESG reporting platforms that extract data from energy invoices — supplier, delivery period, energy usage, and CO₂eq quantity. Developers lack objective data on the energy impact of their infrastructure choices, making it hard to balance speed, cost, and sustainability.
Which web architecture — serverless, classic on-premises, or PWA — consumes the least energy when running an AI-driven application?
Built three variants of the same Flask-based AI app using Gemma 3 12B (server/cloud) and Gemma 3 1B (PWA edge). Measured CPU & GPU energy with Scaphandre, NVIDIA GPU Exporter, Google Cloud Monitoring, and CodeCarbon.
PWA wins per invoice (local compute), serverless scales on-demand via Google Cloud Run, classic infrastructure pays an always-on cost. Results translate to a decision tree for IT architects.
Contact
Looking for a developer who delivers and collaborates well? Let's talk.