Fighting fraud with AI. Restoring the truth in media.
"In a world where anyone can generate a fake image in seconds, I build the AI systems that tell you what's real. The truth in media shouldn't be optional."
I'm an AI Data Engineer at Vaarhaft, where I build data pipelines and ML systems that detect manipulated images, deepfakes, and synthetic content. In a time when AI-generated media is eroding trust in what we see, I work on the other side, giving that trust back.
My work sits at the intersection of computer vision, fraud detection, and AI safety. I design systems that analyze visual media at scale, identifying manipulation patterns that human eyes miss. Every pipeline I build is a line of defense against digital deception.
Before diving into AI engineering, I studied biotechnology at BHT Berlin and Natural Sciences at TU Berlin. The scientific rigor I learned there, hypothesis-driven thinking, data analysis, experimentation, is exactly what makes me a better engineer today.
At 42 Berlin, I taught myself software engineering from the ground up in a peer-to-peer environment with no teachers and no lectures. I founded the AI Club, won hackathons, and built everything from shell interpreters to raycasting engines in C.
I care deeply about AI safety and responsible deployment. The same models we use to fight fraud can themselves be attacked. Robustness, alignment, and accountability aren't abstract concepts to me, they're daily engineering challenges.
AI Data Engineer working on fraud detection and image authenticity verification. Building data pipelines and ML systems that analyze visual media to detect manipulation, deepfakes, and synthetic content, protecting organizations and individuals from digital deception. Connected with CEO Linus Kameni at the 42 Berlin campus, which led to joining the team full-time.
Immersed in one of the world's most innovative coding schools, a peer-to-peer environment with no formal teachers, no lectures, and no tuition. Built projects from the ground up in C, C++, and full-stack web technologies. Founded the AI Club to collaborate on AI projects with peers. Won the Data Science Meets Biology hackathon (in cooperation with Bayer), building an ML model for manufacturing process analysis. Also participated in the AI for All hackathon organized by Anthropic and 42 London.
Monitoring and validation of pharmaceutical production processes. Collected and analyzed structured data for quality assurance, ensuring compliance with industry standards.
Conducted research on micro algae cultivation, gathering and analysing data on growth and CO2 metabolism. Processed experimental data with Python and R for bachelor thesis on evaluating the feasibility of micro algae cultivation on house facades.
Studied biotechnology at BHT (Berliner Hochschule für Technik). Also studied Natural Sciences in the Information Society at TU Berlin, bridging science and technology. Developed rigorous analytical thinking and scientific methodology that translates directly to engineering and data work.
Fraud detection, computer vision, image authenticity analysis, and model evaluation. Building ML pipelines that operate at scale on real-world data.
Systems-level programming in C and C++, backend services in Python. Designing architectures that are robust, efficient, and maintainable.
Full-stack web applications with modern frameworks. Building performant, accessible interfaces and reliable APIs that serve real users.
Scientific methodology applied to engineering problems. Data pipelines, analytics, and the quantitative thinking that comes from a research background.
Deeply engaged with the questions of responsible AI development, alignment research, and ensuring that powerful systems are deployed for beneficial purposes.
Linux environments, containerized deployments, CI/CD pipelines, and version control workflows. Building the infrastructure that keeps software reliable.
An interactive fractal explorer rendering Mandelbrot and Julia sets with infinite zoom, smooth coloring, and multiple palettes.
Try the demo →A Unix shell built from scratch in C. Command parsing, pipes, redirections, environment variables, and built-in commands.
Try the demo →A 3D raycasting engine inspired by Wolfenstein 3D. Real-time first-person rendering of a maze using the DDA algorithm.
Try the demo →Built a machine learning model for performance analysis of a new manufacturing process. Our team won the hackathon, demonstrating how data science can optimize industrial biotech processes.
Read more →Developed AI solutions for charity organisations in an international hackathon organized by Anthropic and 42 London. Collaborated with teams across countries to build tools that make AI accessible for social good.
Read more →Three-day intensive CTF bootcamp at UM6P in Ben Guerir, Morocco. Reverse engineering, threat modeling, and browser exploitation with international teams from 42 Berlin, Heilbronn, Wolfsburg, and 1337 Coding School. Went in with zero CTF experience, walked out with cybersecurity skills and connections across the 42 Network.
Read more →Co-organized and hosted a 4-hour AI-driven coding hackathon at 42 Berlin as co-founder of the AI Club. The winning project, TeamMatch, was a Tinder-style web app matching programmers to projects. Brought the Berlin coding community together for a student-led sprint where creativity meets modern tech.
Read more →The same AI systems we build to detect fraud can themselves be adversarially attacked. Exploring the critical intersection of AI safety principles and real-world fraud detection systems, and why robustness matters as much as accuracy.
Read more →From biotechnology labs to AI-powered fraud detection, going full-time at VAARHAFT. Why non-linear career paths are undervalued in tech, and how every stage of the journey adds a layer.
Read more →What happens when you remove teachers, lectures, and grades from a coding school? Reflections on 42 Berlin's radical model and why learning to learn may be the most important skill in an engineer's toolkit.
Read more →I'm always interested in discussing AI safety, fraud detection, or new opportunities. Whether you have a project in mind or just want to exchange ideas, I'd love to hear from you.