Hi, I'm João Donato
Informatics Security Master's Graduate @ University of Coimbra
About Me
Learn more about my background, skills, and what drives my passion for technology
Years Experience
Projects Completed
Technologies
Hours of 🎾 played
Whoami
I am a recent University of Coimbra graduate with a Master's in Informatics Security, building upon a Bachelor's in Informatics Engineering. My passion and expertise lie at the intersection of AI and security, with a specific focus on enhancing the safety and robustness of Large Language Models (LLMs)
When I'm not coding or messing with LLMs, you can find me listening to podcasts, reading blogs, or playing tennis. I believe in continuous learning and staying updated with the latest technologies and trends.
I thrive in collaborative environments and enjoy working on challenging projects that push the boundaries of LLMs.
What I Do
AI Security
Researching and developing techniques to make LLMs safer.
Cybersecurity
Studying and researching cybersecurity and privacy.
Software Development
Developing applications and websites.
Technologies & Tools
Hover to pause and explore all technologiesAll technologies displayed • Click to learn more
Experience, Education & Certifications
My professional journey, educational background, and professional certifications
Work Experience
Working to protect AI from misuse and ensure that AI systems remain resilient and safe!
Skills:
As part of my master internship, I was awarded a research scholarship in the scope of the AI-SSD project
Technologies:
Skills:
Technologies:
Skills:
Education
Presented a thesis titled 'Benchmarking LLM Robusteness against Prompt-based Adversarial Attacks'
Achievements:
- Average Grade: 19/20
- Published and presented a short abstract paper in the EDCC 2025 conference
Certifications
The AI Security Masterclass is a 4-week course designed for professionals to gain hands-on expertise in AI red-teaming. It covers techniques like prompt injections and defense mechanisms against adversarial attacks.
Credential ID:
gropleeklkProjects
Work in Progress
This section is currently being updated with my latest projects. Check back soon or visit my GitHub for more!
Competitions & Awards
Recognition for technical excellence and competitive achievements
Won 1st place in the AI Red Teaming CTF: [ai_gon3_rogu3], the first AI red teaming CTF from Hack The Box.
Won 6th place in the Agent Red Teaming arena also getting multiple speed-based prizes.
Secured second place the ShiftAppens 2023 hackathon with the queueR project, a tool for helping manage queues.
Won 3rd place in the LLMail-Inject challenge, a competition at bypassing prompt injection defenses in a simulated LLM-integrated email client.
Won 3rd place in the SplxAI's Christmas CTF, a competition at bypassing prompt injection defenses by extracting a secret from the LLM.
Competitions with prizes
Prize Money
Participants Competed
Latest Articles
Sharing insights, tutorials, and thoughts about web development, programming, and technology
Featured Articles
How I defeated all first levels across every agent in Lakera AI's new Agent Breaker challenge - a comprehensive guide to breaking AI defense systems.
My master's thesis on benchmarking LLM robustness against prompt-based adversarial attacks
Short paper presented at EDCC 2025 conference on benchmarking LLM robustness against prompt-based adversarial attacks
Get In Touch
I'm always interested in hearing about new opportunities, projects, or just having a chat about technology