Detect Alzheimer’s disease 20 years early through speech-based algorithms
Alzheimer’s disease and other dementias will affect 1 in 3 people, causing progressive loss of memory – loss of what makes us who we are. No disease modifying treatment currently exists. We now know that the disease progresses silently for decades before symptoms become obvious and a diagnosis is made, which is a major reason why treatments don’t work. But very subtle changes do exist in the silent stages – in episodic memory, executive function, and language – and, with sophisticated enough technology, these changes can be detected in the way someone speaks.
Novoic is a clinical stage biotechnology company developing algorithms to detect neurological diseases such Alzheimer’s disease in their preclinical stages, by analysing audio-linguistic patterns of speech. Building on decades of research, the company is currently testing these algorithms in the clinic, to validate them as software-as-a-medical-device, and bring the first clinically viable speech-based algorithms to market. The company is venture capital funded, working with the Alzheimer’s Drug Discovery Foundation and Gates Ventures to develop a global standard for speech biomarkers, and is the only speech company to have been backed by the UK NHS’ AI organisation to test out its algorithms in 5 common neurological indications.
Founded by Oxford and Cambridge researchers, our team is a reflection of the pioneering nature of our work. Physicians, clinicians, and neuroscientists work side-by-side with machine learning researchers and software engineers, all united by a mission to help patients and their families through early, non-invasive diagnosis.
Why we need you now
Through our ongoing clinical programs, we are collecting what we believe are the best clinical speech datasets for training models to detect neurological disease, in terms of the study design, the speech tasks and the scale of the data. This puts us in a unique position to develop the first speech-based diagnostic tests that can be integrated into the clinic and actually work. We’ve built a great research team but we’re lacking the engineering knowhow and infrastructure we need to maximize the reproducibility, traceability and speed of our research – and help to ensure that our data-handling and algorithm-building remains secure, ethical and compliant with software-as-a-medical-device regulations. We urgently need someone to help us with this now.
What will you actually do
- Architecting and building out MLOps pipelines to maximize the reproducibility, traceability and speed of our research output.
- Building infrastructure for receiving, storing and analysing clinical data in a way that’s secure, ethical and compliant.
- Building CI/CD pipelines for automating code quality assurance and speeding up development for our internal and open-source packages. For example, take a look at our GitHub repos.
- Setting up a framework to log research, deployment of models and experiment tracking in a quality management system to make our work useful for regulators.
- Owning data formatting and data wrangling to make our clinical speech data ML-ready.
- Helping define a great engineering culture and leading engineering hiring.
Who are you
- Strong background in data engineering, including MLOps.
- Excited about being part of a small, quickly growing startup. You want to be a part of our ambitious project. You do not mind the uncertainty that comes with early stage startups.
- Working in healthcare is something that excites you.
- Competitive salary and equity options arrangement.
- 28 days of paid holiday.
- Flexible working arrangements, including hybrid remote work policy.
- A deep sense of meaning, having a significant role in changing the way brain diseases are diagnosed, improving the lives of millions.
- High degree of responsibility and the opportunity to grow into a senior role as the company scales.
- Personal development plan and half a day per week to work on your own projects.