What is a Career Knowledge Graph?
A knowledge graph is a visual representation of interconnected concepts. This Career Knowledge Graph shows my professional journey from 1999 to 2025, mapping over 100 skills, technologies, research areas, and experiences as nodes connected by their relationships.
Unlike a traditional resume or timeline, the knowledge graph reveals how different skills and experiences connect and build upon each other. For example, you can see how my PhD research in robotics (POMCP, particle filters) connects to modern work in logistics optimization and LLM agent evaluation.
Skill Categories
The graph uses 10 color-coded categories to organize skills and concepts:
Core Discipline
Foundational data science skills that underpin everything else. The central hub of the graph.
Academic Research
PhD, MSc, and BSc research in robotics, SLAM, localization, POMDPs, and uncertainty handling.
Logistics
Vehicle routing (VRP), discrete event simulation, matching algorithms, freight optimization from Glovo and Meight.
LLM Agents
RAG, semantic search, evaluation frameworks, structured output, and agent systems work at Meight.
Engineering
Programming languages (Python, Go, C++, Java), databases, cloud infrastructure, and DevOps tools.
Stats & Probability
Bayesian methods, probability theory, A/B testing, hypothesis testing, Monte Carlo methods.
Machine Learning
Scikit-learn, XGBoost, neural networks, clustering, regression, and ML optimization tools.
Languages
Human languages: Dutch (native), English, Spanish, Catalan.
Timeline / Eras
Major career phases: Roots, University of Groningen, UPC Barcelona, Atomian, Glovo, Meight.
Skill Categories
Meta-categories that group related skills: Programming, Databases, Cloud & DevOps, Optimization, AI & ML.
How to Use the Graph
Timeline Slider
Drag the slider at the bottom to travel through time from 1999 to 2025. Watch as nodes appear year by year, showing exactly when each skill or technology entered my toolkit. Press the play button (▶) to automate the timeline animation.
Search
Use the search box in the top-right to find specific technologies or concepts. Matching nodes will remain highlighted while others fade into the background. Clear the search to return to the full view.
Click Nodes
Click any node to zoom in and see its connections. A detail panel appears showing the skill's category, acquisition year, and all connected concepts. Click connected concepts to navigate through the graph. Click the same node again (or the background) to deselect.
Zoom & Pan
Zoom: Use mouse wheel or the +/− buttons in the bottom-right corner.
Pan: Click and drag the background to move around the graph.
Node Sizes
Node size represents importance or depth of experience. Larger nodes (like "Data Science", "Python", "Robotics") are hub concepts I've worked with extensively. Smaller nodes represent specific techniques or tools used within a narrower context.
Connections
Lines between nodes show relationships. Thicker lines represent stronger connections. For example, "Python" connects strongly to "Data Science", while "POMCP" connects to "Monte Carlo" and "PhD Research".
Timeline Highlights
Started programming with QBasic and Turbo Pascal. First exposure to Dutch, English, and the foundations of computing.
BSc in Computing Science (SLAM, GNG networks), MSc in Artificial Intelligence (Global Localization, Particle Filters). Learned Java, C, and Python. Foundation in statistics and probability theory.
PhD in Robotics focused on searching and tracking people using POMDPs, MOMDP, POMCP, and Bayesian methods. Worked with ROS, C++, and humanoid robots. Learned Spanish and Catalan.
Senior R&D Software Developer. Built information extraction systems for medical records. First exposure to SQL, PostgreSQL, and production data systems.
Senior Data Scientist. Built discrete event simulator for delivery networks. Implemented VRP (Vehicle Routing Problem), matching algorithms, and network optimization. Extensive work with experimentation (A/B testing, Switchback testing), AWS, Docker, and Google OR-Tools.
Senior Data Scientist. Designed LLM agent evaluation frameworks (DeepEval, TruLens, LLM-as-a-Judge). Working on freight optimization with tour suggestions, EU regulations (Reg 561/2006), telematics integration. Learning Go, TypeScript, React, and Next.js for full-stack agent development.
Technical Implementation
This knowledge graph uses a force-directed layout algorithm with custom physics simulation. Nodes repel each other while connected nodes attract, creating an organic layout that reveals natural clusters of related concepts. The graph is built with React, SVG, and runs entirely in your browser with no backend.