Source: Image from patent application from The Markov Corporation

Last week we learned about Spero Ventures, a new $100 million investment fund that had been spun out of the Omidyar Network. According to Forbes, the new venture fund has eBay founder Pierre Omidyar as its sole limited partner and counts a company called Markov as one of its first investments.

What exactly the still-stealth Markov does is a mystery, at least until you do a little digging. Markov’s team includes much of the founding team behind the retail point of sale startup Clover Network (including CEO Leonard Speiser), as well as engineers from the likes of Facebook, Linkedin, and Google. While neither Markov’s website nor Crunchbase details how much the company has raised (as of this writing Crunchbase didn’t even have Markov listed), the company subtly lets us know this information via Linkedin, where the company profile says they’ve raised $20 million.

What do they plan to build with that money? Markov’s website uses purposefully vague language about what exactly it does (“The Markov Corporation builds products that can see the world, understand it, and then take an action”), but does say they are building “convolutional neural networks and deep reinforcement learning combined with hardware to make magic.” Their Linkedin page indicates they are building AI that will automate hardware that does “boring work”.

AI. Automation. Neural networks. While that sounds like any number of pitches coming out of silicon valley nowadays, things got interesting for me when I saw how the company’s head of channel sales, Rich Miller, describes Markov: “Revolutionizing a space that hasn’t seen innovations in decades using Machine Learning, Artificial Intelligence, and Computer Vision to change the way food is cooked in commercial applications.”

Based on that description, we can surmise that company is building AI for cooking. In a way that sounds similar to the technology behind June, which uses computer vision and machine learning to help their high-end consumer smart oven cook, only Miller indicates Markov’s target market is commercial kitchens.

All of which what makes what I found next even more interesting: two patent applications from the company’s founders. According to the first patent application entitled “Electronic oven with infrared evaluative control“, the company has developed technology for a control system that utilizes infrared camera sensors to assist in the cooking process. The patent application describes how they plan to use an infrared camera as part of a learning and control system that will more evenly apply heat as compared to a more traditional microwave oven.

This patent application appears to be related to another a patent application from the company founders called “Electronic oven with reflective energy steering“, which describes a way to use RF/microwave energy to more precisely and evenly apply the heat within the cooking chamber. This patent application includes an image that compares and contrasts cooking eggs with a traditional microwave:

To a system (using the technology in the patent) that more evenly applies energy:

Based on these clues it looks like Markov has developed an AI controlled electronic cooking appliance that utilizes infrared vision to gather data as part of a control system that more precisely and evenly applies electromagnetic energy to heat and cook food. While we have no idea if this is all Markov has developed – there might be a lot more they are working on – that they’ve applied for patents for this lets us know it’s what they see as one of their main core technologies.

It’s interesting to note that there has been significant IP built over the past couple years for new solid state cooking systems that use RF energy in the cooking process.  The demo from Miele at IFA showed off an oven that utilizes solid state cooking with technology that applies RF energy with surgical precision.  Early in 2017, the IBEX One, a solid state cooking appliance for pro kitchens, was announced as well with similar technology.

How exactly this Omidyar-invested company’s technology will be used in practice is where things could get interesting, so we’ll be keeping an eye out for further developments.