Postdoc, Deep Learning for Large Language Models and Privacy Translated is involved in the EU DataTools4Heart project ( https://www.datatools4heart.eu/ ), which is a collaborative effort involving sixteen research groups and clinical organizations across Europe. The project aims to advance the diagnosis of cardiological diseases by leveraging cutting-edge machine learning methods and promoting knowledge sharing across languages and countries. To achieve this, we are going to develop a privacy-preserving cardiology data toolbox that includes standardized data ingestion and harmonization tools for a common data model, multilingual natural language processing, federated machine learning, differentially private text generation, and seven language models adapted to the field of cardiology. Additionally, the project will create an open database of synthetic data, CardioSynth, which will be available for further research and development.At Translated, we are actively involved in several challenging tasks within the DataTools4Heart project, such as training large multilingual language models (LLMs) with differential privacy, translating datasets into the seven languages used, synthesizing realistic medical notes, and developing tools to support clinicians in smart cardiovascular risk scoring. We are also interested in publishing our findings in top machine-learning conferences and journals.Through this project, Translated is exploring innovative approaches to various aspects of medical natural language processing to enhance the current state-of-the-art.As a Research Scientist at Translated, you will lead the research efforts for the DataTools4Heart project. You will be a member of Translated's AI team, which focuses on various translation-related products, including expressive speech synthesis, Bayesian data analysis of translation quality data, and internal ML products.You will work within a dynamic research and development group comprising young and experienced professionals based in Rome and Trento, Italy. Remote work may be possible for a limited period of time.In this role, you will:Represent Translated as part of the large international DataTools4Heart project, attending regular consortium meetings and participating in business trips across Europe.Manage and organize tasks related to Translated's contributions to the project.Develop and implement novel approaches to differential privacy for large language models.Assist with the integration of data-driven tools.Preparing and publishing research findings in conferences and journals.Your profile (Desired qualifications) We are seeking a highly qualified candidate with the following profile:Ph.D. in Computer Science, Data Science, or related fields.Strong interest or experience in deep learning and large language models.Ideally, prior experience working with differential privacy and federated learning.Proficient programming skills in PythonPractical knowledge of common machine and deep learning programming frameworks, e.g., PyTorch, Huggingface, Scikit-Learn.Proven ability in managing teams and supervising students.Proactive, goal-oriented, self-starter with excellent time management skills.Benefits and perks Our working environment is both relaxed and intense. We are passionate about our mission, and our work is highly regarded in our industry.Competitive and exciting work environment. You will be surrounded by innovators and experts working at Pi Campus , a venture fund and startup ecosystem. Great environment to grow your skills.We host regular tech and entrepreneurship talks and events, to which you can take part as a Pi Citizen.Work hard and stay fit. In the campus you'll find a gym, a swimming pool, a personal trainer for spinning, TRX and pilates clas ses.What was the hardest piece of work you produced using neural networks? Have you implemented algorithms from recent publications? * How many multi-head attention layers does the encoder of a Transformer (big) have? *#J-18808-Ljbffr
