This deliverable summarizes the work that has been done in the core of T4.2 (“Model Library: Implementation and Recalibration of Adaptive Models”) until M16. This is the first version of a series of deliverables of this task, whose main objective is to provide details on how the developed prediction, prevention and intervention models will be stored and continuously refined or recalibrated based on the feedback gathered from the AI algorithms and user-centric applications. The iHelp Analytic Workbench integrates a model library and a workbench in the iHelp Big Data Platform. The Analytic Workbench enables the “ingestion” of data analytics functions / tasks and their definition in a declarative way to enable the development of adaptive preventive and intervention models for different risks and contributing factors associated with Pancreatic Cancer. The implemented models are stored in a library within the iHelp Big Data platform that offers relevant APIs to support the continuous enrichment and/or adaptation of development models based on the availability and analysis of more (risk, predictions, feedback) data. The analytic workbench also facilitates openness and usability by allowing any actor (e.g. researcher, healthcare professional, data provider, etc.) to develop on-demand adaptive learning models for different risks based on the application of advance AI analytic techniques. This deliverable is being released on M16 of the project and its main aim is to specify the Analytic Workbench along its functionalities. Moreover, the description on how to train and deploy any AI model will be provided so iHelp Clinician partners can train and deploy their own AI models to serve their own needs in the long run, kicking off this way the exploitation activities of the component. An updated and final version of this document will be released on M32.