Algorithm Developer in Mathematical modelling

Do you want your engineering skills to directly shape how athletes understand and improve their movement? Are you excited by the opportunity of turning noisy, high-dimensional data into clean, actionable biomechanical metrics used by professionals and ambitious amateurs alike?

We are looking for an Algorithm Developer in Mathematical modelling to join our Motion Analysis team. You’ll work at the intersection of deep learning and real-world sports science, taking the raw output of our AI models and building the modelling and post-processing pipelines that make the data real. This is where the engineering meets the athlete.

As a member of our team, you will be responsible for the key step between AI prediction and product output. Our deep learning models produce body key points from video. Your job is to turn those noisy, uncertain estimates into robust skeletal fits and reliable biomechanical metrics that our users can trust and act on.

We believe it takes careful post-processing and modelling to turn raw deep learning predictions into data that is real, reliable, and actionable. You will design and develop custom algorithms that handle outlier detection, optimization, and human body model fitting, often as one integrated modelling approach rather than separate steps. You will be part of an agile, collaborative team, contributing to solutions from prototype to production and helping shape our products. You will join a team of six, with diverse backgrounds spanning deep learning, computer vision, software engineering, and mathematical modelling.

We work in a collaborative environment where team members support each other and share responsibility for moving projects forward. We believe that strong solutions emerge through open dialogue across disciplines. We also value a positive and engaging work environment, where curiosity is encouraged and people can explore problems that interest them alongside supportive colleagues.

Your primary responsibilities will include:

• Designing and implementing mathematical models that fit skeletal structures to noisy key point data, balancing input uncertainty with anatomical constraints.

• Developing robust outlier detection and rejection methods that work within the modelling framework, not just as a pre-processing step.

• Building and refining optimization pipelines for post-processing deep learning outputs into accurate biomechanical metrics.

• Conducting data analysis and statistical validation to understand model performance and propose improvements.

• Collaborating closely with our deep learning engineers and software engineers to ensure your models integrate cleanly into the broader product pipeline.

• Exploring and evaluating new approaches from research literature, adapting them to our specific problem domain and data characteristics.