My research sits at the intersection of AI in healthcare, medical data science, and translational biomedical applications. I am interested in building machine learning systems that solve real problems for clinicians, microbiologists, and biomedical researchers, rather than optimizing methods in isolation from practice.

Across projects and application domains, my work is organized around three connected goals:

1. Integrating multimodal health data to support clinically meaningful decisions.
2. Developing trustworthy biomedical AI that experts can understand, evaluate, and use.
3. Releasing open-source software and public data resources that make advanced AI methods easier to adopt.

Research Themes

Multimodal Clinical Decision Support

My current work in microbiology focuses on using machine learning to support surveillance, outbreak detection, organism characterization, and antimicrobial-resistance-related tasks from routine clinical data. I am particularly interested in combining sources such as MALDI-TOF mass spectra, metadata, and expert knowledge in ways that are useful in real diagnostic workflows.

Representative outputs:

• MARISMa: a routine MALDI-TOF MS database from 2018 to 2024 and the associated public dataset
• Automatic antibiotic resistance prediction in Klebsiella pneumoniae based on MALDI-TOF mass spectra
• Automatic Discrimination of Species within the Enterobacter cloacae Complex Using MALDI-TOF MS and Supervised Algorithms
• Automated web-based typing of Clostridioides difficile ribotypes via MALDI-TOF MS
• Automatic surveillance of Escherichia coli bacteriological strains within clinical settings

Trustworthy AI for Biomedical Applications

I work on applied machine learning across multiple health domains, including microbiology, speech, cardiology, and medical imaging. I see this breadth as part of a unified agenda: building AI systems that are useful under real biomedical constraints, especially when data are heterogeneous, limited, noisy, or difficult to interpret.

Representative outputs:

• MARTA: a model for the automatic phonemic grouping of the parkinsonian speech
• Exploring the Power of Photoplethysmogram Matrix for Atrial Fibrillation Detection with Integrated Explainability
• Automatic semantic segmentation of the osseous structures of the paranasal sinuses
• Bayesian automatic screening of pneumonia and lung lesions localization from CT scans

Open-Source Software and Data Resources

An important part of my work is turning research outputs into reusable resources. I care about tools and datasets that help health professionals and biologists use AI more effectively, more transparently, and with less technical overhead.

Representative resources:

• Software & Data
• NeuroVoz dataset
• CT-SCOPE dataset
• KSSHIBA tailored for MALDI-TOF data

Current Directions

• AI for microbiology workflows. I am currently developing an AST application as an open and reusable decision-support tool built around practical microbiology needs.
• AI with domain experts. I want to keep building projects in close collaboration with clinicians, microbiologists, and biomedical scientists so that the resulting models solve the right problem.
• Independent research program. My long-term goal is to lead a group in applied machine learning for health, centered on trustworthy AI, multimodal biomedical data, and open translational software.