The Applied AI Lab
at Grata Health
We study how technology can make people feel understood, supported, and remembered: the things that make care human.
Our research explores how language models can hold emotional context over time, how they can listen in a way that feels steady, and how they can help people stay engaged in recovery when life gets hard.
We work closely with clinicians, researchers, and patients to build systems that make digital care more personal and less mechanical.
Research Areas
Emotional Memory
Teaching models to remember the moments that matter: what someone's working through, what motivates them, and what they've already shared. Conversations build over time instead of starting from scratch.
Empathy Modeling
Developing ways for Luna to pick up on tone, rhythm, and subtle shifts in mood, and respond in ways that feel natural, not scripted.
Trust and Safety
Designing memory and data systems that stay private, encrypted, and specific to each individual.
Recovery-Centered Design
Building and testing everything in real addiction recovery settings, alongside licensed clinicians, to ensure Luna's support complements (not replaces) human care.
Why This Matters
Many people struggle to stay in care because digital systems feel cold or forgetful. Our goal is to build technology that feels reliable: something that listens, adapts, and stays with you. The research behind Luna is what allows that. A model of empathy that grows with every conversation.
Our Team
Luna Research is a small group of engineers, scientists, and clinicians inside Grata Health. We share a belief that empathy can be engineered responsibly: not to replace people, but to help them feel supported between them.