Noah Rubinstein
About Noah Rubinstein
Noah Rubinstein AI Architect at SewerAI
Noah Rubinstein holds the title of AI Architect at SewerAI, where he has played a pivotal role in developing advanced computer vision models. These models are designed to accelerate the inspection process for municipal infrastructure, addressing significant challenges in the sector. His work focuses on leveraging deep learning to process thousands of hours of video footage, ultimately enhancing safety and reducing inspection backlogs for municipalities.
Noah Rubinstein's Work on GPU Utilization
In his role at SewerAI, Noah Rubinstein has significantly contributed to the optimization of GPU utilization. He improved GPU usage from 25% to over 95%, a substantial advancement that has enhanced the efficiency of machine learning processes. This improvement was achieved by scaling batch inference in a cost-effective manner across hundreds of GPUs, allowing for more efficient and economical operations.
Noah Rubinstein's Achievements in Cost Reduction
Noah Rubinstein has also been instrumental in reducing operational costs at SewerAI. By leveraging Anyscale’s Workspaces feature to build distributed Ray applications, he helped achieve a 75% reduction in the total cost of ownership compared to using AWS Batch. This financial efficiency is critical for the scalability and sustainability of infrastructure inspection processes.
Noah Rubinstein's Role in Improving Issue Diagnosis
An important facet of Noah Rubinstein’s work at SewerAI is improving the accuracy of issue diagnosis in municipal infrastructure. His efforts have led to diagnostic performance that exceeds that of expert human inspectors across all defect types. This has been pivotal in elevating the overall safety and reliability of infrastructure services.
Noah Rubinstein's Enhancements in Code Testing Efficiency
In addition to his work on infrastructure and GPU optimization, Noah Rubinstein has enhanced the efficiency of code testing processes. He reduced the time required to test each code change from 10 minutes to nearly instantaneous. This rapid testing capability supports faster development cycles and more responsive innovation in AI-driven infrastructure solutions.