Guust Goossens

Finished building this site. The idea is a GitHub contribution calendar where each cell is a day, and each day can have life commits: things I built, places I went, ideas I chased. Added origin stories going all the way back, wrote proper stories for every entry, set up ImageKit for photos. The goal is for this to actually feel like a life, not a highlight reel :)

Got selected for Null Fellows {}, a European fellows program comparable to Z Fellows. 1 of 5 startup founders from a 30-person cohort. Access to the Stockholm and Helsinki deep tech networks. The program starts in a few weeks; building alongside 5 other founders in Stockholm.

Mistral's worldwide hackathon in Paris. 1,700 applicants, 100 selected. Built a code intelligence tool solo over 24 hours: it indexes any codebase into an interactive knowledge graph with multi-dimensional metadata layers. Added voice input through Voxtral and voice output through ElevenLabs so you can query your codebase while away from a screen. The pitch angle was making coding more human and less physically damaging.

Europe's largest student hackathon. 4,000 applications, 762 teams selected. We won three tracks: 1st in the Google DeepMind Gemini API track (89 teams), 1st in the Lovable track (75 teams), and 3rd in Best Use of Data by Susquehanna/SIG (152 teams, the quant finance one). Built an agent forum pattern where agents explored connected data sources, posted findings to a shared knowledge base, and a structuring agent organized everything into a navigable hierarchy. The three wins span genuinely different evaluation criteria: data pipeline quality, AI integration, and product polish. Walked away with €50K in DeepMind credits, €1k in Lovable credits and some cash prizes.

IC Hack at Imperial College London, the UK's biggest student hackathon. Built ERLA, a recursive research agent that spawns branching sub-agents to explore scientific literature exhaustively. When context gets too large, a managing agent autonomously splits the search into branches. Added a hallucination detection pipeline called HaluGate that hit 95%+ groundedness on validated summaries. Placed 2nd overall in the Emergent Behaviour track.

First hackathon ever. Carnegie Mellon in Pittsburgh. Built a working demo, pitched it, and met a ton of cool builders and investors from across the US. The product was fine, but the people were the best part. Left with a clear sense that I wanted to do a lot more of these.

Got selected for a hackathon at Carnegie Mellon in Pittsburgh. My first hackathon. I like the idea of hackathons because the constraints force you to just build. No hiding behind roadmaps and planning docs. Either you ship something in a day or two or you don't.

Kicked off a new project. Can't say much about it yet. The shape of it is taking up most of my building time though, and I'm enjoying the problem.

End of first semester at Polytechnique. My approach is always the same: spend most of the semester on client work and real building, then fully switch to exam mode when Parkinson's law kicks in. It keeps working, so I keep doing it.

Built a RAG system for ClearTax. Ingests their internal docs and knowledge base, answers queries with citations. Early testing showed about 60% less time spent on research. The interesting part: at this scale, the quality of the vector search matters way more than which LLM you use.

Kaggle competition on churn prediction. Tried a hybrid Random Forest + XGBoost ensemble, then vibecoded some neural network approaches: CNNs on tabular data, LSTMs for sequential patterns. Also played around with survival analysis for longer-term predictions. In the end, the classical ensemble just worked better on this dataset. Sometimes that's how it goes.

Hit €45k+ in cumulative consulting revenue over 12 months, while studying at Polytechnique. Mostly AI engineering projects; they have better margins and I genuinely enjoy them more. Each client project teaches me something new, and the revenue funds whatever I want to build next.

Made a Next.js app connected to the Mistral SDK. You give it any topic and it generates a structured learning plan: phases, knowledge items, quizzes, all through recursive structured output. Basically a personal tutor that helps you go deep instead of just skimming the surface.

Started the MSc X-HEC at École Polytechnique. Machine Learning, Statistics, Algorithmic Optimization. Having a strong institution on the CV means I can spend less energy worrying about credentials and more energy actually building things.

Set up my own management company to formalize the freelance web development and AI engineering work I'd been doing. Having a proper structure made it easier to take on bigger clients and invoice properly. The agency became the vehicle for everything that followed.

Done! Four years at KU Leuven. Honestly a lot of the most valuable things I learned happened outside the curriculum. Through Plenti, the agency, and all the rabbit holes. But the degree gave me the structure to explore all of that, and I'm grateful for it.

Went to London with my sister and visited the 'Electric Dreams' exhibition at Tate Modern. It traces the history of computing and art: plotters, early generative systems, the first time machines made something beautiful. Bought the catalogue. Left feeling like the people who built the early internet cared more about aesthetics than we give them credit for.

Final exam period. During this stretch I also got the news that I was accepted into X-HEC Data Science & AI for Business. I can completely redirect my focus when the deadline gets close enough. Parkinson's law on steroids.

Plenti was making money, but the HR management side made scaling really difficult. My co-founder and I decided to wrap it up. No regrets. Teaching myself to code for Plenti, with AI as my tutor, is probably the single best thing I've done for my career so far.

Ran the half marathon in Ghent with my mom. 21km through the city, started at 7am. We both got the medal. It was genuinely one of the nicest mornings I've had in a while.

Joined Dirac, a student society focused on value investing and pitching. Nice way to stay close to the fundamentals: unit economics, moats, the things that still matter when trends fade.

Traveled the east coast of Mexico with friends. The car rental companies were either sold out or way too expensive, so we ended up borrowing a car from the neighbor at our first Airbnb. Completely saved the trip. Sometimes the best logistics are the ones you improvise.

Exchange semester at ESSEC in Paris. Arrived drained after a breakup and some family stuff; the first weeks were genuinely hard. But slowly things turned around. Replaced my student job with freelance web development work, kept building Plenti remotely, and fell in love with the city. Paris ended up becoming my favorite place.

Co-founded Plenti, a platform connecting people who need help (tutoring, gardening, babysitting) with people who can provide it. We needed someone technical, so I taught myself full-stack development from scratch. Next.js, React Native, Tailwind, PostgreSQL, all learned on the job. Most of it through conversations with Claude, which opened my eyes to what's possible with AI. If you combine it with active thinking, it's an amazing all-knowing one-on-one tutor that never loses his chill. We started with €3,000, got to €1,500/month passive income, and served 60+ clients in the first month. One of the steepest learning curves of my life.

New York with my cheapest friends. We spent a total of €1,000 for 11 days by finding an insane flights plus hotel plus breakfast deal in New Jersey, then doing a bunch of metro hopping and living on $1 pizza slices. Some of my favorite photos come from this trip.

Joined Beta Sigma Club, an international student society for quantitative finance. Ended up becoming Head of Operations pretty quickly. Got to coordinate things across chapters in Leuven, Paris, Munich, and London. It was the first time I felt like I was around people who were into the same weird intersection of finance and tech.

Collaborated with a developer from Lebanon to turn my manual NFT minting token strategy into an algorithm. Around the same time I got into buying NFT sniping bots, on-chain trading tools, and went way deeper into the conceptual reality of crypto: how blockchains actually work, MEV, liquidity dynamics. The rabbit hole kept going and I kept following it.

Turned 18, so I could finally be independent with my crypto passions. Was working at a fancy restaurant, and the day I turned 18 I opened a bank account and started putting most of my money into NFTs. Found a systematic price movement in NFT minting tokens on Solana: they were given to active community members of certain collections, and their prices always spiked in the hours before minting. Lots of people were selling them way too low, so you could make good money swinging them. Failed most of my exams but ended up making me and my friends over €3,000 in crypto trades. Taught me that a degree is not a necessary ingredient, or even a valid ingredient, for doing cool things in life.

Started my bachelor's at KU Leuven. Business & Information Systems Engineering. A lot of my actual learning happened outside the curriculum though: rabbit holes into algorithmic trading, crypto economics, how platforms scale. The degree gave me structure, the rabbit holes gave me direction.

At 17, while Bitcoin went from 11,000 to a new all-time high of 22,000, I finally got my father to let me invest €100 in crypto. Instantly turned to leveraged trading on Binance. Watched my portfolio go from 100 to 550 in a very roller coaster way. It taught me a lot about technical analysis, crypto trading, and opened my eyes to quant trading, which was really underdeveloped at the time for crypto.

At 12-13 I got really into Rocket League, a video game where you play football with cars. Eventually discovered you could trade in-game items, and slowly got addicted to the trading instead of the football part. Spent months reading demand through websites and Excel sheets, watching tutorials, timing trades, building up inventory. After getting scammed a few times, I sold my items for real cash to a Russian guy through PayPal. €200. Felt like a fortune, bought a lot of candy with it. Taught me the concept of alpha: the investor who knows his market and assets best picks out the best opportunities. Also taught the pros and cons of trading in the wild west.

Trading Pokémon cards on the playground. Got weirdly into figuring out which cards were underpriced and finding the right person to trade with. Probably the first time I got excited about how markets work, even if the market was just a bunch of 10-year-olds flexing shiny toys. Also turned out to be a growing market :)