Product Science: AI-powered APM tool for Mobile App Performance is founded by four siblings: Maria, Anna, David, and Daniil Libermans who have been building startups together since 2006. The siblings’ expertise came from computer vision, graphics, image recognition, and AR. In 2016 Snapchat’s parent company Snap acquired its startup Kernel AR. In Snap the siblings led the development of 3D Bitmoji Avatars – so far the most popular digital avatar in the world, with over 1 billion avatars created and over 250 million daily active users. At Snap Inc they championed solving performance issues for the Android app – a successful project that led to the revamping of the user and revenue growth in 2018. That was when they realized that existing performance and observability tools were ineffective and decided to reinvent the APM (application performance management) industry for the mobile-first world.
Product Science is AI-powered software that analyzes application code in a multi-threaded environment, finds suboptimality in the code execution that affects performance, and helps every engineer to improve code on the go, without involving experts and using user data.
When a mobile app is running, its code runs on dozens of threads simultaneously. These execution paths create an extremely complex 4-dimensional maze. To allow engineers to navigate within this maze and analyze only the important parts of it, Product Science employs AI to identify a single, most critical execution path which is responsible for a satisfactory accomplishment of a specified user action.
The Product Science toolset is a combination of two key components. The first component is the dynamic automatic code instrumentation via AI-powered plugins added to the build process. In addition to instrumentation, the engineers get access to the multi-threaded code profiling tool highlighting critical functions and frameworks that impact user experience, revealing the root cause of the problem. By replacing manual instrumentation and embedding right into the build processes, Product Science enables anyone to identify the causes of app performance issues.
Product Science is the first in the industry, an AI-powered, pre-production APM tool for Mobile App Performance.
Product Science has addressed a longstanding issue in mobile development by providing developers with increased visibility into the execution of their code on devices. Traditional native tools like Perfetto offered limited functionality and required extensive manual instrumentation and logging efforts. Product Science’s solution works alongside existing application performance monitoring (APM) and observability tools, such as Sentry and DataDog. These tools are proficient at identifying issues but cannot identify root causes on the client side of mobile applications.
By enabling engineers to see their apps for the first time, Product Science helps to identify and solve issues before shipping the app to the users. The customers of Product Science see improved developer productivity and app speed. The solution’s automated instrumentation enables the collection of data in a more cost-effective and meaningful manner, enhancing the value of observability. Compared to other mobile performance optimization efforts, which focus on single-threaded regression analysis or narrow use cases, Product Science is taking a unique position in the field. This approach can be beneficial to any mobile engineer, regardless of their specific use case.
And there’s come Product Science: an AI-powered APM tool for Mobile App Performance!
Top engineering teams of Fortune 500 companies in sectors such as banking, sharing economy, and social media use Product Science tech.
Since the latest funding round, the Product Science team has released its product with a completely new UI with a flow that helps engineers break down performance engineering processes.
With the new UI, Product Science optimizes critical mobile app flows to increase the engagement and retention impact of our tool.
It recently introduced another breakthrough feature, which synchronizes video with tracing. Now engineers and product designers can see the video recording of their app synchronized with the profiler data recorded on any mobile device. It has never been done before in profiling, and it makes navigating through traces so natural, as you scrub through a video recording and dive into the code executed behind the scenes.
Product Science’s mission is to eliminate delays caused by software inefficiency for people worldwide. Time is humanity’s most valuable non-renewable resource, and Product Science’s goal is to help individuals and organizations save time and improve efficiency through innovative technology solutions. As AI powers most of the technological breakthroughs nowadays, it is now used to optimize software to its fullest potential, empowering people to focus on what matters most.
The team’s AI engineers are working on datasets and mapping the functions that provide insights into the code’s optimal performance. This enables Product Science’s plugin to filter out functions that a native profiler would normally record. Reducing noise in large data sets is crucial to optimize performance and prevent overwhelming.
Another feature on the team’s roadmap will enable engineers to identify and optimize performance issues while coding in their preferred IDEs.
Coatue, K5 Global, and Slow Ventures backed Product Science.
Performance greatly impacts user acquisition and product metrics, so technical, data and product leaders should be aware of this impact. Product Science provides tools that drastically reduce the impact of performance issues on business metrics.
For example, the tool helped Saturn, an emerging social media application, to identify critical issues affecting user acquisition and retention. The observability tools measured Saturn’s app start as having an average of 0.6 seconds of wait time which contradicted the engineer’s observations and users’ complaints. Product Science showed that the app responded to user actions only after 4 seconds of wait time. The tool identified the precise sections of the code that require attention. Product Science provided actionable insights to the engineers, resulting in a decrease in wait time from 4 to 1.8 seconds in the first month and 0.7 seconds in the second month. This allowed the engineers to concentrate their major resources on shipping features while improving app performance.