Visualization Startup Brytlyt Combines AI, GPUs, PostgresSQL
The world may seem like it has slowed down in the midst of a pandemic, but technologies like real-time data analytics keep accelerating as one startup after another comes up with new ways to crunch numbers and leverage the results.
Among them is U.K. analytics and visualization startup Brytlyt, which announced a $4 million funding round this week. Investors include Amadeus Capital Partners and Finch Capital, bringing Brytlyt’s total funding to a modest $6 million.
The visualization company specializes in helping telecommunications carriers and others sift through huge data sets—up to 1 terabyte—generated by millions of handsets. Once organized, those data are used to create maps, charts and graphs for predictive analytics applications.
The London-based startup claims its platform is among the first AI- and GPU-based analytics platforms built on PostgresSQL. That combination is said to yield sub-second responses to multiple tables joining billions of records. Brytlyt further claims benchmark testing showed its SQL database runs 30 times faster than other GPU-based platforms, 300 times quicker than in-memory databases and 1,000 times faster than legacy systems.
As the combination of GPU and deep learning help accelerate data analytics, Brytlyt said it has added proprietary AI software to the mix to deliver sub-second visual data analytics.
“Analytics and data are growing exponentially, and it’s clear the future of processing must lie in GPUs, with their enhanced capacity for data crunching and deep learning,” said Richard Heyns, Brytlyt’s founder and CEO.
“We are bridging the gap between the old SQL world and the new and exciting world of AI, enabling companies to bring their infrastructure up to date and extract value from their data in real time,” Heyns added.
The company’s flagship product, the SpotLyt visualization tool, is designed to handle complex queries on billions of rows of data. Charts, graphics and maps are delivered in milliseconds.
Among the reasons for the mapping speed-up is the shift in rendering from the client to the database side. That framework means only the image is shipped, not the underlying data.
Besides visualizations, the tool also includes an SQL editor to produce the coding that interacts with a PostgresSQL database. The tool also is linked directly to Jyputer notebooks.
Along with telcos, Brytlyt works with energy, financial, healthcare, logistics and retail customers in areas ranging from fraud detection and risk management to network performance optimization and supply chain management.
The company said it is also collaborating with researchers tracking the spread of the novel coronavirus.