What is MACH?
MACH (Multilingual AI Conversational agent delivering information on HIV) is a conversational agent/chatbot that provides reliable lifestyle information and general information about HIV. It uses a large language model (LLM) fine-tuned specifically for HIV-related topics, ensuring responses are accurate, relevant, and sensitive to users' needs.
In addition, MACH supports multiple languages through machine translation, making it accessible to a diverse range of users worldwide.
We deploy robust evaluation frameworks and protocols designed to verify that MACH functions as intended while remaining accurate, safe, fair, equitable, actionable, and evidence-based.
Why We Created MACH
Access to accurate HIV information should not depend on where you live, the language you speak, or whether you feel comfortable asking for help. Yet stigma, misinformation, and language barriers continue to make it difficult for many people to find the answers they need.
MACH was created to change that — combining the expertise of clinicians, people with lived experience, and AI developers to create a tool that offers trusted information with empathy and respect. Our goal is simple: to empower people to make informed, confident decisions about their health.
Our Values
Compassion
We believe every question deserves kindness and respect.
Inclusivity
MACH supports users of all backgrounds, genders, and sexualities — in many languages.
Privacy
Your conversations are encrypted and never stored.
Accuracy
Information is guided by trusted sources such as the NHS, WHO, UNAIDS, and BHIVA.
Our Approach to Privacy and Security
Your privacy matters. MACH is built with security at its core:
- No personal data is collected or stored.
- All chats are end-to-end encrypted.
- Fully compliant with GDPR and NHS Digital information governance standards.
We believe that trust begins with transparency — and that means protecting every user’s confidentiality.
How MACH Ensures Accuracy
MACH ensures accuracy through a combination of rigorous evaluation, continuous validation, and evidence-based design. Its algorithms and decision processes are regularly tested against verified datasets to confirm precision and reliability in outputs.
Quality assurance protocols — including cross-validation, peer review, and real-time performance monitoring — are embedded throughout the development and deployment phases. MACH also integrates user feedback loops and automated error detection systems to identify and correct discrepancies promptly.
The Team Behind MACH
MACH is developed by a team based at Mildmay Hospital, bringing together expertise in HIV care, clinical practice, and digital innovation. The team is committed to using technology to reduce stigma, improve understanding, and make reliable health information accessible to everyone.
Divyan Moodley
Project Lead & Clinical Safety OfficerMBChB · MSc Computer Science with AI
Divyan is a resident doctor in HIV at Mildmay Hospital and holds an MSc in Computer Science with AI. He conceptualised MACH to give every patient — regardless of language or background — access to clear, stigma-free information about their health. As Chief Investigator of MACH's clinical validation studies, he is pioneering the evaluation of large language models for patient education, including the assessment of commercial AI systems for patient-facing applications.
To create a world where everyone, regardless of language, can access trustworthy health information and feel empowered to care for their well-being.
Our Vision
Contact Us
We welcome collaboration with clinicians, researchers, and community organisations.
info@machhealth.co.uk