Diogo Antunes
AI Scientist & Entrepreneur
Specialized in Machine Learning and Deep Learning with expertise in transformer architectures, reinforcement learning, and enterprise AI systems. Building the future of intelligent software at the intersection of research and practical applications.
Pioneering Value-Driven AI Systems
At 23, I'm redesigning artificial intelligence to be autonomous, safe, and value-aligned. Through Vaisys Inc., we're building the next generation of AI that moves beyond conversation to meaningful action—always with humans in control.
The Value-Driven Approach
As an AI Scientist and entrepreneur, I believe artificial intelligence should serve humanity with clear values: transparency, collaboration, and responsible innovation. My work focuses on moving beyond the limitations of current language models to create systems that can autonomously execute complex tasks while maintaining human oversight.
Through deep expertise in transformer architectures, reinforcement learning, and advanced optimization techniques, I'm building AI systems that understand not just what to do, but why it matters. This perspective drives innovation that creates impactful, trustworthy technology.
My vision extends beyond technical excellence—I'm an entrepreneur who sees AI as a tool for positive societal transformation. Every system I design prioritizes safety, value-alignment, and the principle that AI should amplify human capability, not replace human judgment.
Language → Agentic Action Model
Moving AI beyond conversation to safe, autonomous execution with humans in the decision loop.
Value-Driven Framework
Transparency, collaboration, and responsible innovation guide every AI system we build.
Enterprise-Grade Impact
Scaling businesses faster while ensuring innovators stay in control and society shares the rewards.
Principled Guardrails
Pairing frontier research with robust safety measures for trustworthy AI deployment.
“I'm not just building AI systems—I'm reimagining how artificial intelligence can serve humanity with clear values, autonomous capability, and unwavering commitment to keeping humans in control. The future of AI is value-driven, and that future starts today.”
— Diogo Antunes, Founder & CEO of Vaisys Inc.
Research & Publications
Advancing AI research through innovative methodologies, from published work to cutting-edge private research driving the future of value-aligned AI systems
Academic Thesis
Scan Optimization with AI
Bachelor's Thesis • Instituto Politécnico do Cávado e do Ave • July 2024
End-to-end machine learning implementation covering data engineering, model training, and production deployment. Applies Longformer architecture for sequence classification with custom tokenization for hierarchical data structures. The work addresses the complete ML pipeline from data preprocessing through optimized inference, demonstrating practical approaches to structural pattern recognition tasks.
Internal Research & Development
Language → Agentic Action Models (LAM): Beyond Conversational AI
Vaisys Internal Research • Proprietary • 2025
Proprietary research into next-generation AI systems that move beyond language models to autonomous action execution. This foundational work explores novel architectures for safe, value-aligned AI agents capable of complex reasoning and decision-making in enterprise environments. Includes breakthrough techniques in human-AI collaboration and autonomous task decomposition.
Constitutional AI for Enterprise Safety Guardrails
Vaisys Internal Research • Proprietary • 2025
Developing robust safety frameworks for autonomous AI systems in enterprise environments. This research establishes constitutional principles for AI behavior, creating fail-safe mechanisms and ethical decision boundaries for agents operating in business-critical contexts. Includes novel approaches to AI transparency and explainability.
Technical Expertise
Specialized in cutting-edge AI technologies and production-ready machine learning systems
AI & Machine Learning
Expert in state-of-the-art AI technologies with focus on transformer architectures, reinforcement learning, and advanced model optimization techniques.
Advanced AI Techniques
Specialized in cutting-edge AI methodologies including human preference learning, multi-modal systems, and scalable model architectures.
Data & Validation
Advanced expertise in data validation, quality assurance, and building robust data pipelines for AI systems at scale.
Production AI Systems
End-to-end AI system development from research to production deployment with emphasis on scalability and reliability.
Experience
Building AI systems from research to production at scale
Vaisys
Lisbon, Portugal • Geographically independent
CEO, Founder
Full-time
Established Vaisys with a vision to create AI that genuinely benefits society, prioritizes ethical practices, privacy, and—most importantly—builds solutions with a purpose over mere reasoning or technical prowess.
Principal Engineer
Full-time
Developing and refining solutions, reinforcing Vaisys's mission to deliver tangible, value-driven AI systems under a trustworthy and private platform.
Artificial Intelligence Scientist
Full-time
Researching specialized transformer neural networks designed for unique use cases. Managing datasets of over 50 billion tokens for 2 billion parameter models.
DefensePoint
Lisbon, Portugal • Remote
AI Tech Lead Consultant
Contract
Leading AI technology initiatives as a specialized consultant, focusing on advanced machine learning solutions for security-critical applications.
AI Tech Lead
Full-time
Led AI technology development initiatives, designing advanced machine learning solutions for Identity and Access Management systems.
Checkmarx
Braga, Portugal • Hybrid
Software, Machine Learning Engineer
Full-time
Solely responsible for finetuning state-of-the-art open-source autoregressive models for internal use, incorporating proprietary data using QLoRa methods.
Software, Machine Learning Engineer
Internship
Implemented transformers based approach for sequence classification. Involved in training models from scratch and engineering tokenization process using Longformer architecture.
Projects
Open-source contributions and AI-powered applications
Housify AI Platform
Comprehensive generative AI system with distributed inference capabilities for multiple AI models including LLava, Stable Diffusion, and Pix2Pix.
Tech Stack
TuxTech E-commerce
Full-featured e-commerce platform built with Django, featuring user management, payment processing, and inventory systems.
Tech Stack
Computer Vision Traffic Detection
Real-time traffic signal detection system using OpenCV and C++, implementing advanced image processing algorithms.
Tech Stack
Education & Certifications
Continuous learning through formal education and professional certifications
Bachelor of Science in Software Engineering
Polytechnic Institute of Cavado and Ave
Grade: 80% • Thesis: "Scan Optimization With AI"
Relevant Coursework: Algorithms & Data Structures, Linear Algebra, Statistics & Probability, Calculus, Machine Learning, Computer Vision, Natural Language Processing, Artificial Intelligence, Software Architecture
Professional Certifications
Fundamentals of Deep Learning
NVIDIA Deep Learning Institute • 2022
Issued by: NVIDIA
Hands-on course covering neural network training, optimization techniques, and deployment strategies using NVIDIA's deep learning frameworks and GPU acceleration.
Fundamentals of Accelerated Computing with CUDA C/C++
NVIDIA Deep Learning Institute • 2022
Issued by: NVIDIA
Comprehensive program covering CUDA programming, parallel computing fundamentals, and GPU acceleration techniques for high-performance computing applications.
Fundamentals of Accelerated Data Science with RAPIDS
NVIDIA Deep Learning Institute • 2022
Issued by: NVIDIA
Advanced data science acceleration using RAPIDS framework, covering GPU-accelerated data processing, machine learning, and analytics workflows.
Disaster Risk Monitoring Using Satellite Imagery
NVIDIA Deep Learning Institute • 2022
Issued by: NVIDIA
Specialized course on applying computer vision and deep learning techniques to satellite imagery for disaster monitoring and risk assessment applications.
Google Advanced Data Analytics
Google • 2023
Issued by: Google Career Certificates
Comprehensive data analytics program covering advanced statistical analysis, machine learning, and data visualization techniques using Google tools and methodologies.
Security Champion
Checkmarx • 2024
Issued by: Checkmarx Ltd.
Professional certification recognizing advanced expertise in application security, secure code practices, and security-first development methodologies. Demonstrates proficiency in identifying vulnerabilities and implementing security best practices.
Pioneering Value-Driven AI Systems that harness enterprise-grade agentic artificial intelligence to boost business performance while benefiting everyone
A Delaware-based startup shifting AI from Language → Agentic Action Model (LAM), moving beyond conversation to safe, autonomous execution. Where enterprises scale faster, innovators stay in control, and society shares the rewards.
Autonomy
End-to-end execution with zero babysitting. AI that works as a tireless collaborator.
Reliability
Self-monitoring, self-healing systems with human-override capabilities. Technology people and enterprises can trust.
Velocity
Hours of work collapsed into seconds. Cutting downtime, shrinking dev cycles, and amplifying human creativity at scale.
Human-Centric
Surfaces decisions and keeps people in control. Guided by transparency, collaboration, and responsible innovation.
Large Action Models in Practice
Agentic AI that runs, reasons and acts for enterprises 24/7
Continuous Autonomous Ops
Watches logs, fixes incidents, writes post-mortemsSelf-Maintaining Code
Profiles, refactors, tests and commits automaticallyDecision Intelligence
Researches, models and drafts reports for leadersCross-Domain Flexibility
From data engineering to robotic control—same model, new roleSelf-Evolving
Adjusts from every run through memoryAlways Human-in-Loop
Autonomous execution with human oversight and controlMission & Vision
Vaisys is guided by a value-focused ethos: transparency, collaboration, and responsible innovation. We build technology people and enterprises can trust, creating AI that genuinely benefits society through ethical practices and purpose-driven solutions.
Our vision is to harness advanced, enterprise-grade agentic artificial intelligence that boosts business performance while benefiting everyone—ensuring artificial intelligence amplifies human capability rather than replacing human judgment.
Learn More About Vaisys
Let's Build the Future
Ready to collaborate on innovative AI projects, research opportunities, and transformative solutions
Open to discussing AI research, enterprise solutions, startup opportunities, and projects that push the boundaries of machine learning
Location
Lisbon Metropolitan Area
Portugal & Remote
Available for in-person meetings in PortugalAvailability
Open for Networking
Consulting & Collaborations
Ready for professional networking and partnershipsConnect & Follow
Ready to Start Something Amazing?
Whether you're looking to push the boundaries of AI research, build enterprise solutions, or explore new frontiers in machine learning, I'm here to collaborate and create impactful innovations.