PEOPLE. NATURE. KNOWLEDGE.
We bring science and artificial intelligence together to create new possibilities in health, education and sustainability.
From knowledge to transformation.
NATURE + INTELLIGENCE
Universidade
Niterói, Rio de Janeiro
ABOUT BIOD-AI
We are the Research Group for Artificial Intelligence Engineering for Biodiversity, Sustainability & Health, affiliated with the Biosystems Engineering Graduate Programme (PGEB) at Fluminense Federal University.
We connect different fields of knowledge to develop products, processes and computational solutions. Real-world challenges are our starting point. Our aim is to expand autonomy, strengthen research and build more sustainable futures.
OUR PROJECT ECOSYSTEM
Each project connects scientific knowledge to a human, environmental or social need. Explore our initiatives and discover what we are building.
10 projects shown.
Communication and autonomy
Everyone has something to say.
Assistive technology to support communication for people with aphasia and other difficulties with expression. We bring artificial intelligence and accessible interfaces into the everyday lives of individuals, families and professionals to support autonomy, participation and human connection.
Gaze-based interaction
When a gaze opens up possibilities.
Speech, writing and computer interaction for people who rely primarily on eye movements. NeuroVox explores eye tracking and neural networks to turn gaze into the selection of letters, words and commands, accommodating different movement abilities.
A connected scientific resource
Images in context. Research with purpose.
An image repository for neurology and neuroscience research, linked to scientific evidence and reference resources. NeuroLab brings structure, context and knowledge together to help researchers explore materials relevant to their studies.
Integrated research and innovation
More science. Less fragmentation.
An integrated, one-stop solution for research and innovation in neurology and neuroscience. NeuroDesk brings data, artificial intelligence, neural networks and scientific analysis together in an environment designed to support researchers throughout their work.
Personalised education
Every student is unique. So is their learning.
Digital learning twins to understand educational journeys and adapt assessments to each individual. LearnerTwin combines artificial intelligence and educational curation to support teachers, respect differences and turn educational data into more personalised attention.
Regeneration and sustainability
Intelligence to restore possibilities to the soil.
Bayesian modelling and digital twins to support the regeneration of degraded soils. Inspired by Amazonian Dark Earth, AmazonSoil Twin brings data and scenarios together to inform the assessment of bioinputs, remineralisers and land management strategies, with explicit attention to uncertainty.
Health data and scenarios
Connecting data to reveal possibilities.
An intelligent platform for governance, data integration, Bayesian prediction and digital-twin simulation. HealthTwin supports scenario development for health research and planning, combining structured information with uncertainty analysis.
Technology for care
Care without taking away independence.
Intelligent monitoring to identify possible falls among older adults and notify designated contacts. Zellus AI explores computer vision and alert workflows to connect the detection of an incident with the people who can help.
Intelligence for markets
Making decisions means understanding uncertainty.
Bayesian neural networks applied to the analysis of financial and commodity markets. BayesCore AI connects modelling, scenarios and data visualisation to support research into risk, volatility and time-series behaviour.
CONCEPTUAL VISUAL10 / BIOD-AI
Brain health research
Research today to open new paths tomorrow.
Artificial intelligence research into neurodegenerative diseases. Neuro Sentinel investigates how computational models can support the exploration of data, patterns and new scientific questions.
Projects are at different stages of research and development. Talk to the team to learn about the scope and conditions of use of each initiative.
BIOD-AI NEUROTECH
Our technology division develops products, solutions and resources to support research in neurology and neuroscience.
We bring together assistive communication, gaze-based interaction, scientific image organisation and integrated research environments. We connect researchers, professionals and individuals to turn important questions into new possibilities.
CONNECTIONS THAT CREATE KNOWLEDGEExplore the division’s initiatives
Our focus is research, innovation and assistive technology. Our scientific platforms are not presented as substitutes for diagnosis, clinical reporting or clinical decision-making.
OUR TEAM AND EXPERTISE
From modelling to experimentation, and from data processing to social applications, our complementary expertise supports scientific projects and studies with depth, collaboration and responsibility.
Neural networks, machine learning and the representation of uncertainty.
Modelling and processing with supercomputers, including Santos Dumont at LNCC.
Quantum computing research through the partnership with IBM.
Mathematical and computational methods and analytical support for scientific projects and studies.
Research into materials, processes and applications to develop solutions.
Research into land management, bioinputs and the recovery of degraded soils.
Technologies to support research, accessibility and innovation in brain health.
Solutions focused on autonomy, education and expanding opportunities.
Modelling and analysis to support decisions and public policy development.
Have a scientific question, a technology challenge or an idea for collaboration?
OUR HOME
Our laboratory is at NAB, on UFF’s Praia Vermelha Campus. A meeting place for research, knowledge and collaboration.
NAB is a UFF research centre focused on biomass studies and water management. This academic environment brings together Biod-AI’s research initiatives.
Rua Professor Edmundo March, s/nGet in touch to arrange a visit and receive directions for accessing the laboratory.
GET IN TOUCH
Research, development, partnerships or new ideas: tell us what drives you. Let’s explore the possibilities together.
contato@biodai.com.br
LET’S CONNECT IDEAS
INFORMATION AND TRANSPARENCY
This is an institutional Biod-AI page. Project descriptions and the team’s expertise were prepared from information supplied by the group.
Do not include sensitive information, examinations, medical records or other health data in your initial contact message. For collaborations involving research data, ask the team about appropriate channels and procedures.
This page’s code does not include analytics tools, advertising pixels or its own cookies. The hosting provider may retain technical access logs according to the server administrator’s configuration.
The NAB logo and photograph are requested from the official NAB/UFF gallery. The location map is loaded from OpenStreetMap when the visitor approaches the section. These requests share technical connection information, such as the IP address, with the respective providers. Location links open Google Maps only when chosen by the visitor. The page does not request GPS location or device location permission.
When external resources are unavailable, a text-based NAB identifier and a location reference image are displayed. These alternatives are not an official logo, a laboratory photograph or a street map.
On the first visit, this page estimates the country associated with the IP address to choose the language: Portuguese for Brazil and English for other countries. It does not request GPS, precise location or device location permission.
The lookup uses the Country API (api.country.is) and, when needed, GeoJS (get.geojs.io). These providers receive the IP address and technical connection information. The site code uses only the returned country; it does not store the IP address, transmit form content or use the lookup for advertising.
A manual PT or EN choice is saved in browser local storage and takes priority over automatic detection. The country may be retained in session storage for up to 30 minutes, without the IP address. The “Automatic language” option removes the manual preference. If the lookup fails, the default is English, with manual switching always available.
VPNs, proxies and corporate networks may make the estimated country differ from the visitor’s physical location. Country detection is not proof of location.
The initiatives are at different stages of research and development. The neurology and neuroscience research platforms are not presented as tools for diagnosis, clinical reporting or clinical decision-making. Assistive technologies, simulations and analytical resources must be assessed in their application context and with the involvement of responsible professionals.
Descriptions of BayesCore AI are not investment advice or a promise of returns. Agronomic applications of AmazonSoil Twin depend on local data, validation and professional responsibility. Zellus AI emergency integrations depend on specific protocols and deployment.
The Portuguese and English project presentation artwork was supplied by Biod-AI and is displayed according to the selected language. The “About Biod-AI” image is an AI-generated conceptual illustration; it is not a photograph of the laboratory or team. Institutional logos and names are preserved in their original form.
NAB — Centre for Biomass Studies and Water Management.
References consulted on 13 September 2026. Contact the team to confirm laboratory access and the current status of the initiatives.
Communication and autonomy
Everyone has something to say.
AphasAI places communication at the centre of care. The project develops digital resources to support people who face difficulties expressing needs, choices and feelings, with attention to accessibility and family involvement.
Its ecosystem includes assistive communication interfaces and opportunities for gaze-based interaction. Collaboration with professionals and research centres guides improvements to the experience and the assessment of its use in different contexts.
Gaze-based interaction
When a gaze opens up possibilities.
NeuroVox explores a form of communication in which gaze becomes the control mechanism. Calibration and on-screen selection enable users to compose messages and explore digital interactions without a conventional keyboard or mouse.
Development combines computer vision and neural models, with attention to lighting conditions, head movements and each user’s individual characteristics. The aim is to expand opportunities for expression and access to digital environments.
A connected scientific resource
Images in context. Research with purpose.
NeuroLab is an initiative to bring together images of scientific interest and the context that helps researchers interpret them. Rather than simply collecting files, it connects visual material with references, structure and useful research resources.
The NeuroTech division is developing this environment to support information access and the work of neurology and neuroscience researchers. Use of the collections must comply with the relevant access permissions, licences and data protection requirements.
Integrated research and innovation
More science. Less fragmentation.
NeuroDesk addresses the need to reduce fragmentation between research tools and stages. It aims to bring computational resources together to support data exploration, the application of artificial intelligence and studies in neurology and neuroscience.
Neural networks, analytical methods and scientific workflow organisation are part of this integrated vision. The aim is to provide a research and innovation environment for researchers and physicians engaged in scientific work, connecting knowledge and tools.
Personalised education
Every student is unique. So is their learning.
LearnerTwin builds an individual view of each learning journey to support the creation of assessments tailored to the student. The technology is designed to support teachers, who remain responsible for educational guidance and reviewing activities.
The proposal connects students, teachers, schools and education groups around more contextualised assessment. Human curation and classroom use analysis guide improvements to the system.
Regeneration and sustainability
Intelligence to restore possibilities to the soil.
AmazonSoil Twin brings together soil knowledge, computational modelling and Bayesian artificial intelligence. Its reference to Amazonian Dark Earth guides research into the regeneration, fertilisation and management of degraded areas.
The aim is to compare possible interventions and represent uncertainty before expanding field applications. Bioinputs, remineralisers and management combinations are analysed in a process connecting real data, experimentation and technical oversight.
Health data and scenarios
Connecting data to reveal possibilities.
HealthTwin brings together four complementary areas: governance, integration, Bayesian prediction and digital-twin simulation. The aim is to turn fragmented data into a more structured foundation for investigating scenarios and supporting health planning.
Rather than presenting predictions as certainties, the Bayesian approach incorporates uncertainty into analytical work. The platform is designed to build knowledge and assess possibilities according to data quality and the context of each project.
Technology for care
Care without taking away independence.
Zellus AI is being developed to recognise signs of possible falls in images and send alerts to designated contacts. Its focus is supporting care networks and helping identify situations that may require attention.
Response workflows take the context of the event and deployment conditions into account. The project includes the development of mechanisms for escalating care, subject to the protocols and integrations actually available.
Intelligence for markets
Making decisions means understanding uncertainty.
BayesCore AI develops computational intelligence to analyse time series and scenarios in financial and commodity markets. The approach combines Bayesian neural networks and visualisation tools to represent possibilities and uncertainties.
The project connects modelling research with analysis and monitoring environments. Its focus is helping users understand data and risks, recognising that historical results and projections do not guarantee future market behaviour.
Research content and analytical support. This does not constitute investment advice or a promise of returns.
Brain health research
Research today to open new paths tomorrow.
Neuro Sentinel is a research initiative exploring artificial intelligence in the study of neurodegenerative diseases. The work brings computational modelling together with scientific questions related to brain health.
Its purpose is to support the exploration of data and patterns that may inform new research hypotheses in neurology and neuroscience. Interpreting the results requires validation, context and the involvement of specialist professionals.
