We use open algorithm design techniques to create deployable probability models.
Using a combination of text parsing and network analysis we can construct identity consolidation functions.
Using classical and modern techniques we can create segmentation groups based on a variety of customer dimensions.
Designing and training deep learning models that quickly deploy in lightweight environments.
models — supervised models and deep learning approaches to work out the odds of something happening.
machine learning research — we implement findings in new research and get practical results.
segmentation — customer investigation, understanding and storytelling.
economics — industry and country investigation. customised insight report generation.
data matching — customer identity consolidation, single customer view matching and processing.
infrastructure — deployable model serving tools (cloud or on-site) and open source analytics tool stack.
consumer — we extract data from hundreds of systems to build models to understand people.
finance — managing credit models and predicting future usage patterns.
health — primarily in the health insurance space specialising in customer consolidation and matching.
social/startup — pro-bono quality analytics for bootstrapped companies and not-for-profits.
workload split 40/60 between R&D and consulting with impactful industries worldwide.
partnered with universities to introduce new deep learning methods to industry.
registered B-Corp with a singular dedication mission to ‘recondition the human condition’.