Anomaly detection system

The realm of energy storage systems presented an intricate puzzle - detecting unpredictable anomalies in sensor readings. Coupled with the critical need for accuracy and reducing false positive alarms, our team was tasked with navigating vast volumes of data, ensuring system predictability, and developing a user-friendly interface that simplifies complex data outcomes. Regular systems consist of 100 000s of sensors with COGR 20-30% per year giving unprecedented pressure on commissioning and operations staff. This is a new market and companies underestimate the number and complexity of problems they will experience in next 5-10 years with installed systems.

I served as the product manager for this project, overseeing team coordination, collecting technical requirements, and prioritizing them. I was involved in every phase of the product, from ideation to testing. After evaluating the initial MVP, we identified a new set of stakeholders and discovered numerous features that necessitated expert insights. A distinctive aspect of this product is its tailored approach to a niche market, addressing a yet unresolved issue as the ESS market undergoes significant transitions and robust growth.

The fruits of our labor were rewarding. Post system implementation, the efficiency gains were evident, with a significant reduction in the time required for commissioning works by over 30%. Furthermore, our team's dedication to A/B testing and MVP performance evaluations ensured that the solution was not only innovative but also robust and reliable. We achieved substantial results in understanding system behavior and increasing overall accuracy of system. User trainings further solidified our commitment to stakeholder engagement, arming them with the knowledge and confidence to optimize system performance.