Company Mission StatementeMoodie is a digital mental health start-up based in Edinburgh. Our mission is to combat the mental health epidemic unfolding around the world by ensuring that people have access to affordable and reliable mental health tools. We do this by pairing an individual’s self-reported symptoms with digital biomarker data in combination with AI-driven algorithms to determine their current problems and progress while engaging with digitally delivered therapeutics. Our solutions are aimed at giving individuals the tools they need, as early as possible, to effectively manage their mental health - thus preventing the deterioration of milder symptoms into clinical disorders. The mounting tide of mental health problems in our society is a key issue of our modern age and eMoodie is dedicated to helping flatten the mental health curve.Product OvervieweMoodie has two products: a research platform called eMoodie Labs and a mental health app called eMoodie Minds. The current job overview would focus on the latter. eMoodie Minds is based on a unique approach to screening for mental health symptoms such that digital interventions can be targeted to the symptoms individuals are actually experiencing rather than providing generic advice - which is what standard digital mental health tools do. eMoodie Minds MVP is now ready and its initial version focuses on addressing employee stress and burnout amidst the increasing shift to remote working patterns brought about by the pandemic.Typically, digital screening for mental health problems is achieved through self-report questionnaires, but there are many limitations to this approach including biases in one’s memory when reporting on emotionally salient events, and the fact that regular self-reported entries can become burdensome. To address this, eMoodie wants to focus on more objectively collected data. We plan to leverage vocal and facial expression data (e.g. through voice and video recordings), in combination with location and ECG data collected from smartwatches, to build Deep Learning AI models which can predict behavioural markers (e.g. sleep problems, affect, and anxiety) of mental health problems.AI has greatly advanced diagnostics in the medical treatment of physical health problems, but has simply not been applied to mental health with the same rigour. eMoodie’s goal is to advance digital mental health screening and diagnostics so as to mitigate the harms caused by misdiagnosis and the lack of appropriate support that is rife in our health system.Job overviewWe are looking for a highly creative Data Scientist to research and develop the Deep Learning models using the different modalities described above. The ideal candidate is passionate about machine learning and research, and has previous experience building Deep Learning models. Prior experience with the analysis of voice/audio data will be prioritized during the hiring process.You will work with the eMoodie team which will include clinical psychologists who will help to frame and evaluate the screening results, a data engineer, who will be responsible for building the data pipelines, and software engineers, who will integrate the developed models into the mental health app. We expect you to take the lead and be able to work independently on researching the best approaches for processing and analysing the data from the various different sources. A primordial aspect of the approach to be used is that it needs to be done in a way which preserves the users’ privacy.You will also have the opportunity to be listed as a co-author in any subsequent peer-reviewed publications arising from the project. ResponsibilitiesWork closely with the eMoodie team to identify metrics for evaluating the Deep Learning models, taking into account various trade-offs such as costs of wrong predictions etc.Work closely with software engineers and data engineers on the team to devise a technical architecture which puts security and privacy at the forefront.Review existing relevant published work and research existing systems which can be leveraged/re-used.Work independently to develop multimodal Deep Learning modelsWork closely with clinical psychologists to ensure that the output of the models can be used for recommending targeted interventions (interpretability of models)Skills and qualificationsProven experience as a data scientist; manipulating data sets and building statistical models, knowledge of a variety of machine learning techniques (clustering, decision tree learning etc.) and their real-world advantages/drawbacks.Experience working with audio and video data formats and familiarity with deep learning techniques.While we are open to different data science tools, experience working with the Python programming language and relevant libraries would be advantageous. Ability to work independently and be proactive.Excellent written and verbal communication skills, and ability to work as part of a dynamic multidisciplinary team.A drive to learn and master new technologies and techniques.A sincere interest in improving the lives of individuals who suffer with mental health problems.(Bonus) Has past experience analysing voice/audio data.(Bonus) Has a Master’s or PHD in Statistics, Mathematics, Computer Science or another quantitative field.(Bonus) Familiarity with privacy-preserving AI techniques (e.g. OpenMined).(Bonus) Experience with data visualisation and data analytics.(Bonus) Past experience working in HealthTech.Please submit a cover letter and CV.