Company and Use Case Poly Perception, a Belgium-based startup, focuses on delivering advanced AI-powered solutions to the waste industry. Their flagship product is an intelligent analyzer that monitors and optimizes waste sorting processes. By using cameras and machine learning, their system can detect, classify, and estimate the weight of materials in real-time, providing valuable data to help recycling and sorting facilities improve efficiency and achieve higher levels of waste purity.

Performance Problems to Solve Before Timescale, Poly Perception faced challenges in processing and storing large amounts of time-series data efficiently. Their AI system needed to analyze massive amounts of waste sorting data in near real-time, which was initially hindered by slow processing speeds and insufficient data retention policies. The inability to analyze the waste streams continuously created inefficiencies, with only limited snapshots of the sorting process available for evaluation.

Performance Gains Unlocked TimescaleDB revolutionized Poly Perception’s data management. With features like continuous aggregates and efficient time-series data storage, the company now offers real-time waste monitoring, enabling facilities to fine-tune their sorting processes. The transition to Timescale helped Poly Perception reduce latency and store data more efficiently, ultimately enhancing the accuracy and speed of their AI-based waste classification system. As a result, their customers now enjoy operational improvements, increased throughput, and better sorting performance metrics.