Challenge

PSYKHE is an early stage startup that uses AI and psychology to recommend consumer products. Anabel, the founder and CEO, had been managing a small team of data scientists and engineers for the last year, and despite some success was struggling to get to product launch. The lead data scientist had just announced he was leaving and Anabel got in touch with Ortom for some help. She wanted some assurance that all of the work done to date would not be wasted and some help on future direction. Data and machine learning were central to her plans for the company and she wanted assurance she was on the right track.

Solution

We worked together to build and execute an ML and data strategy. This involved a review of systems, code and data, and detailed interviews with the data science and engineering teams. The review revealed that while a lot of good work had been completed, the data science team were getting bogged down in infrastructure and data pipeline issues. Because they were spending a lot of time on work they were not trained for, significant technical debt was accumulating and morale was low. We recommended they hire somebody with data engineering experience to solve pipeline and infrastructure issues and we helped them to build and deploy new features in their recommendation engine.

Result

We helped PSYKHE hire a new data scientist and get some temporary help with infrastructure. We also helped in the handover and documentation process to make sure all the knowledge in the company was not lost. We build a model to incorporate individual user feedback into their recommendation system. The system was implemented and worked exactly as specified. The new system allowed thousands of users to receive personalised product recommendations based on their individual feedback.

Conclusion

The model was handed over to the new team and is now part of the product. PSYKHE received a new funding round on the back of this work.