AiThority Interview Series With Orr Danon, CEO and Co-Founder at Hailo

At Hailo, our technology is based on structure rather than rules, which is far more native to Deep Learning and vastly more efficient.

Know My Company

Tell us about your journey into the intelligent tech industry. What galvanized you to join Hailo Technologies?

I joined this field almost by accident, as I come from a very different background. I was introduced to this exciting new world by Avi Baum, who is my partner and the company’s CTO at the time, he served as CTO of Texas Instruments’s IoT division. Within a very short time, two things became abundantly clear: the huge opportunity which Deep Learning offers, and the unbelievable gap in computer architecture which prevents this potential from being unleashed.

What does it take to start and succeed in a Deep Tech startup ecosystem?

There are many paths one can take, but I would say that there are two main directions: The first would be “specialization”, which entails focusing on a specific tech which you have worked on intensively (whether in the industrial or the academic sphere). In this case, you would build an improved product based on your ability to gather a highly experienced team in this field. The alternative path is more focused on integrating ideas from different fields to create novel systems. In these circumstances, the challenge is to gather highly skilled people from many fields and then concurrently build both a company culture and a product architecture which can fuse all of these diverse ideas into a single coherent solution.

How do you prepare for an AI-driven world as a business leader?

As a company which focuses on the foundational infrastructure for Deep Learning at the edge, we closely monitor various industries to see which will be the first to adopt the technology. We then build solid business cases, which in turn will drive demand for mass deployment. Due to the fact that this is a fast-changing landscape, accurate analysis depends on both technical and business variables. So best practice is to maintain a constant dialogue with innovative players in all of these prospective markets.

What are the foundational tenets of your Deep Learning platform? How does it work to drive AI to the edge of its potential?

Hailo has re-designed the entire technology stack all the way up to the frameworks used to describe Neural Networks. Our technology is based on structure rather than rules, which is far more native to Deep Learning and vastly more efficient.

Hailo’s re-creation of computer processing shatters the barriers of traditional computing architecture, enabling local computation of deep-learning neural networks in edge devices such as autonomous vehicles, drones, or cameras. Hailo’s innovative breakthrough will enable the edge device to employ intensive deep-learning technology to allow mass deployment of intelligent capabilities within a reasonable power, area and cost budget.

What does your ‘Ideal Customer’ Profile look like? Which new geographies are you currently targeting?

A key target market is the current Automotive industry (Tier 1 and OEM), but that is only one of several areas of interest.

The Hailo processor, which is expected to be rolled out in the first half of 2019, is designated for incorporation in a wide range of edge devices including drones, smart home appliances such as personal assistants, smart cameras and smart TVs, IoT, AR and VR platforms, wearables and security products.

We are currently cooperating with several of the largest companies in the automotive industry and other fields, which are examining the possibility of incorporating our technology.

What is the state of ‘AI in Embedded Device’ in 2018-2020?

The need for Artificial Intelligence at the edge is substantial and demand is steadily increasing. The missing piece of the puzzle is the ability to create edge devices that can handle massive amounts of data independently in a data center. Hailo can serve as the enabler in these cases. Once we have a processor that can handle the heavy computation of deep learning in edge devices, even more possibilities will open up. Just think of an old phone with a cable as opposed to a light smartphone you carry on you. This is an unprecedented breakthrough that will enable new developments we haven’t even thought of yet.

What are the major challenges in adoption of AI for embedded devices? How does Hailo Technologies help customers overcome these challenges?

Let’s take the automotive industry as an example: ECUs in autonomous vehicles cannot keep up with the computing load for real-time decision-making and must rely on power-hungry, cumbersome and expensive general-purpose solutions to process the deep learning algorithms for making instantaneous decisions. Hailo’s innovative processor technology design will deliver data center performance while slashing costs and power requirements.

The Crystal Gaze

What AI and machine learning start-ups and labs are you keenly following?

I generally follow all the leading research institutes and large ML unicorns.

Which technologies within AI and Computing are you interested in?

All advancements relating artificial neural networks. Unsupervised, partially supervised networks, training with limited data sets, to name a few.

As an AI leader, what industries you think would be fastest to adopting AI/ML with smooth efficiency? What are the new emerging markets for AI technology markets?

Cameras and camera-equipped devices; these kinds of products are ubiquitous and feature savvy, thus, differentiation based on scene understanding is of the essence

What’s your smartest work-related shortcut or productivity hack?

Just do it!

Tag the one person in the industry whose answers to these questions you would love to read:

Michio Kaku

Thank you, Orr! That was fun and hope to see you back on AiThority soon.