The new agentic AI that you can
unders
VISION
Building human-centric collaborative agents to solve complex problems
Such agents materialized in the form of our first agent Nook who beat human champions at the game of Bridge in 2022. This feat was achieved by NukkAI’s unique and multidisciplinary team of research scientists and engineers employing a wide range of paradigms, from statistical machine learning to knowledge representation and reasoning.
Over the years, NukkAI’s vision for Agentic AI has been refined. Driven by this, we are now building a new form of AI: a multi-paradigm and knowledge-driven AI that places human-AI interaction at its heart.
Why multi-paradigm?
We regard LLM as just one tool within a broader toolkit of technologies with complementary strength. For example, symbolic AI, one of NukkAI’s specialties, provides a robust framework for logical reasoning and planning. By harnessing the advantages of these technologies, NukkAI is leading the revolution towards all-round AI agents.
Why knowledge-driven?
Knowledge is the key to scalable and sustainable AI development and deployment. With structured domain knowledge incorporated, our agents achieve higher performance while consuming less energy and data. They can also be deployed efficiently in various contexts when powered by different knowledge, making them more flexible and adaptive.
Why human-centric?
Trust between humans and AI is built on transparency and mutual understanding. Such confidence can only be obtained by keeping humans fully informed of the decision-making of an interpretable and explainable AI. NukkAI’s philosophy also emphasizes a collaborative partnership between humans and AI, thus leveraging the strengths of both. While working in concert, our agents provide explanations of their decisions to humans, allowing them to offer further guidance and to eventually make the final call. This close-knit relation ensures that AI remains a force for good.
PLATFORM
Collaborative Toolbox
NukkAI is the leader of next generation agentic AI. We empower companies to solve their most complex problems while ensuring humans remain in control.
WE BUILD SOLUTIONS THAT ARE
Collaborative
You can easily communicate with NukkAI’s AI agents, adjust their parameters, and make more informed decisions.
Adaptable
Composed from hybrid technologies and different AI paradigms, our AI tools can be deployed to various complex problems.
Sustainable
By drawing on human knowledge, NukkAI’s agentic toolbox requires less energy and data than other AI engines.
LEVERAGING MULTI-PARADIGM AGENTS
Optimization Agent
Preference Elicitation Agent
Business Experts
Constraints Manager Agent
LLM- Based Extraction Agent
Explainer Agent
Forecast Agent
Orchestrator Agent
Information Analysis Agent
Reporting Agent
Geospatial Monitoring Agent
Knowledge Modelling Agent
Profiling Agent
Core Agents
Strategic Agents
USE CASES
Heterogeneous Data Fusion
We develop heterogeneous data aggregation and ontologies to allow high-level queries and automate report creation at different levels of abstraction, also leveraging AI to discover new patterns. This technology, which will be tested during a NATO war game, is fully relevant for the Defense, Cybersecurity, and Finance sectors.
- Cutting-edged data modelling tools
- Natural language collaboration
- AI-enhanced crisis management system
- Low-maintenance service
Collaborative Planning
We support major companies in complex planning challenges that involve a vast number of combinations and constraints, while providing transparency and regulatory compliance. Among the many benefits for our clients: cost reduction, simplification of the highly complex process for crew pairing, improvement of productivity, enhanced human-machine interaction to keep people in control of operations, and faster iterations.
- Unlimited constraints updates
- Adaptive user experience
- Natural language reporting
- Specific needs adaptability
Real-Time Recommendation
We leverage cognitive science, symbolic machine learning, and LLMs to design adaptive learning agents. These context-aware agents provide users with reliable recommendations and explainable feedback.
- Unexpected events handling
- Agentic workflow automation
- Intelligent document processing
- Enhanced decision-making
TEAM
Jean-Baptiste
Co-Founder & CEO
Véronique
Co-Founder & Head of Research
Swann
Chief Technical Officer
Nouredine
Product Owner
Jean-Pierre
DPO & Project Manager
Mathilde
Chief of Staff
David
Administrative Manager
Junkang
Research Scientist & R&D Coordinator
Malik
Research Scientist
Dimitri
Front-End Developer
Camille
PhD Student
Marc
Research Scientist
Kevin
Research Engineer
Colin
Research Engineer
Danlin
Research Engineer
Annabelle
Business Developer
Nicolas
Marketing & Communication Apprentice
Jean-François
Front-End Developer
Henry
Expert Research Scientist
Gildas
Business Developer
We have broken through an AI barrier: the game of Bridge
In March 2022, NukkAI has hit a major milestone in AI by outperforming 8 bridge World champions during a Bridge challenge.
Until then, Bridge was AI’s unconquered challenge. A step up from simpler games with complete information (Chess & Go), the game of Bridge is a more complex game with teamwork and incomplete information, a perfect testbed for AI of the new generation.
TECHNOLOGY
We rely on the most advanced AI technologies
The originality of our approach is based on its hybrid nature (combination of several AI paradigms), making it possible to process incomplete data in a probabilistic universe while providing explanations to humans. We mix open-source libraries and proprietary algorithms to create the most advanced hybrid AI toolbox.
Deep Learning
Inductive Logic Programming
Signal Processing
Multi-Agent System
Active Learning
Cognitive & Behavioral Science
Abstraction Layer
Genetic Algorithms
Planning
Time Series Analysis
Natural Language Processing
Reinforcement Learning
Ontology Management
Tree search
Advanced Expertise
Proprietary
Our latest publications
“Revelations: A Decidable Class of POMDP with Omega-Regular Objectives”. In 39th Annual AAAI Conference on Artificial Intelligence (AAAI), 2025, Philadelphia, USA.
Read more“The Complexity of Pure Maxmin in Extensive-Form Games”. In Journal of Artificial Intelligence Research (JAIR), Volume 82, 2025.
Read more“Multiannotator supervised learning and common explanations in first order logic.” PhD Thesis defended in November 2024.
Read moreThey
support us
NukkAI is proud to be recognized by outstanding scientists from across academics, industry, and business.
