Simplex HIMES, a certified EHR provider serving clinics across the UAE, needed AI capabilities without transferring patient data. We built an anonymization middleware that processes EHR data inside a TEE
Secure your patients’ health data while implementing AI
Run AI on sensitive patient records without leaking data to the open world. Created by Yma in cooperation with global and UAE champions
The data that would make AI genuinely useful
is exactly the data that cannot be shared
The data that would make AI genuinely useful is exactly the data that cannot be shared
Yma builds Confidential computing
Confidential
Confidential computing
for healthcare
Confidential
computing for
healthcare
We provide cryptographic and operational proof that sensitive healthcare data stays protected while AI work is performed inside controlled environments
Anonymization before AI processing Reduce exposure of patient identities before records are used for inference, fine-tuning, or retrieval tasks
Controlled execution inside TEEs* Keep sensitive data inside attested environments designed for diagnosable and governable processing
Built for regulated healthcare operations Support private AI execution where clinics, EHR systems, and regional partners need stronger privacy controls and clear operational boundaries
* Trusted Execution Environment
Controlled execution model
Protected data flow for healthcare AI
Sensitive records are transformed, processed, and returned inside governed infrastructure
Clinic / EHR data
Trusted Execution Environment
- LLM-based anonymization
- LLM & AI Agents
- Verification
Safe outputs only
In partnership with 
Proven in production
We fine-tuned MedGemma 27B on 120,000+ real doctor-patient dialogues inside a confidential computing environment. The results surprised even us
Benefits from Yma Confidential
Safe access to real data and controlled execution are what make medical AI practical in the first place
Real healthcare data matters
Medical AI only becomes useful when it is grounded in real workflows, real records, and the messiness of actual care delivery
Anonymization is operational, not cosmetic
Protecting patient identities is part of how healthcare organizations keep trust, stay compliant, and open the door to safer data collaboration
Controlled failure boundaries matter
Healthcare AI needs diagnosable, governable infrastructure so failures can be understood, contained, and improved over time
User-cases
Clinics and hospitals
Run diagnostics support, triage support, and treatment-assistance workflows on patient data without redesigning the full operating environment
Pharmaceutical and research teams
Access anonymized clinical datasets for research and model work where real-world data quality matters but privacy still governs the process
EHR providers
Add AI capabilities to certified platforms without moving sensitive data into uncontrolled channels or undermining user trust
Multi-institution collaboration
Support shared model development across institutions while each party keeps data protected inside its own controlled environment
Let’s talk about your data requirements
Whether you are a clinic, EHR provider, or research institution, we will work with you on the right setup for confidential AI execution and protected healthcare data processing
Thank you!
We'll get in touch
with you shortly.