The Challenge: A Europe-based MedTech start-up, which offers medical-grade cardiac home monitoring, is willing to utilise AI to achieve ECG-level performance from a low-cost wearable that is available to the patient any time they need it. The main idea is to detect heart-related medical conditions before they become serious issues. The company is planning for a class 2A certification.Our Approach: As a first step we conducted a thorough analysis of the current ECG signal processing system and algorithms. Our experienced data scientists and ML engineers created an automated tool for testing the existing solution with various internal and external medical data sets. Furthermore, we identified areas for improvements and extensions of the scope of detectable medical conditions. We advised on the certification process and prepared the AI-related documentation for the class 2A certification.Results: Our team performed analysis and recommendations in less than 8 weeks, with actionable advice on the class 2A certification leading to time reductions of the application process, as well as improvement recommendations and execution.

Challenge: The client was a start-up solving the problem of managing numerous processes within a recruitment department. Hiring teams have data from a variety of sources, which are difficult to process and analyze. Our task was to structure and develop a data structure and create an AI tool for screening candidates. Approach: Our work started with a research and development phase. A core task was to develop a solution architecture which meets the requirements of the AI models but of the applications which collected the data already. We identified an AWS structure combined with other ETL tools to be able to work with different formats from multiple sources.  Once the data structure and pipelines were set, we used semantic analysis with NLTK preprocessing to extract the necessary information for each candidate. The ML part was further developed with Keras, TensorFlow and SKLearn. The output presented a structured database of candidates and sorted them into predefined groups and factors.Key Metrics: As a result of the project, applicants in the system are automatically shortlisted and hiring teams need only allocate strategic efforts to review the screening results which proved to have up to 83% more accuracy. The solutions achieved their goal to optimize the recruitment process and reduce the time invested in a manual review of the CVs of candidates by half and thus increase the productivity of a recruitment team using the application.

Challenge: The client is a scale-up which provides health systems and doctors with a digital-first, high-touch solution with innovative technology. The key challenge was to upgrade and grow a platform that is already working with hundreds of thousands of patients and health practitioners.Team:  Python and ReactJS developersSolution: The solution facilitates a standardized workflow, creating a repeatable care plan to provide the same level of care for each patient population and clinical condition. It includes multiple functions and interactions, in complex roles from patients through hospitals, doctors, and pharmacies to the healthcare fund overseeing and managing the whole system. The software has multiple integrations with healthcare devices, which collect and analyse data from patients and provide them to their healthcare specialists.Outcomes: Our team is successfully handling the project and continues to be a trusted partner of the management.