Latent AI, a startup born of SRI International, makes it easy to run AI workloads at the edge by dynamically managing workloads as needed.
Using its proprietary compression and compilation process, Latent AI promises to compress library files 10x and run them with 5x lower latency than other systems, all using less power thanks to its new adaptive AI technology, which the company is rolling out today as part of its appearance in the ProWellTech Disrupt Battlefield competition.
Founded by CEO Jags Kandasamy and CTO Sek Chai, the company has already raised a $ 6.5 million seed round led by Steve Jurvetson of Future Ventures and followed by Autotech Ventures .
Before starting Latent AI, Kandasamy sold his previous startup OtoSense to Analog Devices (in addition to handling HPE mid-range security business before then). OtoSense used data from sound and vibration sensors for predictive maintenance use cases. Prior to its sale, the airline worked with the likes of Delta Airlines and Airbus.
In a sense, Latent AI collects part of this work and marries it with the IP of SRI International .
"With OtoSense, I had already done some board work," Kandasamy said. "We had moved the audio recognition part out of the cloud. We did the learning in the cloud, but the recognition was done on the edge device and we had to quickly convert and remove it. Our account in the first few months got us moving that way. . You couldn't stream data over LTE or 3G for too long. "
At SRI, Chai worked on a project that looked at how to best manage power for flying objects where, if you have a single power source, the system could intelligently allocate resources to power flight or run. onboard processing workloads, primarily for surveillance, then switch between them as needed. Most of the time, in a surveillance use case, nothing happens. And while that's the case, there's no need to calculate every frame you see.
"We took it and turned it into a tool and platform so we could apply it to all kinds of use cases, from voice to vision, from segmentation to time series," explained Kandasamy.
What's important to note here is that the company offers the various components of what it calls the Latent AI Efficient Inference Platform (LEIP) as standalone modules or as a fully integrated system. The compressor and compiler are the first two of these and what the company is launching today is LEIP Adapt, the part of the system that manages the Kandasamy dynamic AI workloads described above.
In practical terms, the LEIP Adapt use case is that your battery powered smart doorbell, for example, can run in low power mode for a long time, waiting for something to happen. Then when someone comes to your door, the camera wakes up to run a larger model, perhaps even on the doorbell base station connected to power, to perform image recognition. And if a whole group of people gets to one (which is unlikely right now, but maybe next year, after the pandemic is under control), the system can offload the workload to the cloud as needed.
Kandasamy tells me that the interest in technology has been "huge". Given his previous experience and SRI International's network, it's perhaps unsurprising that Latent AI is garnering a lot of interest from the auto industry, but Kandasamy also noted that the company is working with consumer companies, including a camera and manufacturer. of hearing aids. .
The company is also working with a major telecom company that is looking at Latent AI as part of its AI orchestration platform and a large CDN provider to help them run AI workloads on a JavaScript backend.
