Our Publications
Our approach

A safety-first and hybrid AI approach

Modular and Controlled AI Systems

Instead of relying on a single black-box model, our specialized NLP tools form a modular AI system leveraging a verified structured database of drug information. This ensures precision and control in healthcare applications.
Explore Our Hybrid AI System

Multilingual Clinical NER

Our award-winning paper, "Multilingual Clinical NER: Translation or Cross-lingual Transfer?", explores methods for clinical NER in multiple languages without annotated data. Our ClinicalNER model, available on Huggingface, demonstrates high performance and versatility.
Read the Full Article
Open Research Commitment

Contributing to Global NLP Understanding

Posos is committed to open research, publishing in international conferences, and open-sourcing our code and models. Our ClinicalNER model has garnered significant recognition and thousands of downloads, reflecting our dedication to transparency and collaboration.
Access Our Open Source Models
Our difference

Aligning Generative AI

Building Safer, Specialized LLMs

Utilizing our structured, multilingual database, we create safer and specialized LLMs for clinical applications. Our proprietary data, currently undergoing clinical trials, enhances our ability to align LLMs with safety and precision.

Cutting-Edge AI Safety Techniques

We leverage advanced methods like Reinforcement Learning from Human Feedback (RLHF) and Direct Preference Optimization (DPO) to drive our research into safer AI models for healthcare. Posos is at the forefront of developing AI prioritizing clinical safety and effectiveness.
Join Us

Shape the Future of AI in Healthcare

Posos invites AI and healthcare experts to join our mission. Collaborate with us to develop AI solutions that are groundbreaking and safe, enhancing healthcare delivery and patient outcomes.