Staff AI Product Manager at Sword Health, leading conversational AI experiences in healthcare.
Leading the Phoenix AI team, developing a conversational AI assistant that integrates across Sword Health's healthcare products ecosystem. Phoenix AI serves as the conversational interface for members using Thrive (AI-assisted physiotherapy), Bloom (pelvic therapy), Move (screenless movement tracking), and Mind (mental health solutions).
Led AI product development at the Center for Responsible AI, a €78M initiative uniting 10 startups, 8 research centers, and 5 industry leaders in Portugal to develop 21 next-generation AI products grounded in principles of social responsibility.
Product Lead for "Halo - Silently Controlled Language Generation", enabling easier communication for those who have lost the capacity of speech. Led collaboration with top national and international research groups including CMU. Also led "Translation for High Risk Content" and "Cultural-Aware Multilingual Customer Service" products.
HALO demo at Casa da Música, Porto
HALO demo with patient at Champalimaud Foundation
Led the roadmap and experiments for data and AI solutions across 3 products in the Unbabel Labs Innovation team. Responsible for collaboration between Labs and Research teams, bringing Responsible AI principles into Labs products, and managing external academic research partnerships.
Contributed to European Projects such as MAIA and UTTER. Co-supervised several PhD and master students.
Part of the small team that developed COMET, the reference state-of-the-art Machine Translation Metric with more than 750 citations. Co-authored COMETinho, which was awarded Best Paper Award at EAMT22.
Research focused on motivation and effort-based decision-making in the general population and children with ADHD, using novel experimental tasks and computational modeling approaches. Graduated with Distinction and Honors by unanimity.
Research on satisfaction metrics vs. behavioral outcomes
AI personality design in therapeutic contexts
Communication preferences in AI-assisted physical therapy