top of page
  • Writer's pictureKathleen Lineberger

Trailblazers: Meet Jim Shea, CEO of DeepSig

We partner with Trailblazers, founders whose innovative ideas are reimagining their industries. This series aims to highlight disruptive companies and the leaders who built them.


 


How did you get started with DeepSig?

I’ve been in the defense industry for over 35 years and had just completed a turnaround of DRS Signal Solutions, a technology company in Germantown, MD and I was ready to get involved in another startup. I was introduced to Tim O’Shea by Charles Clancy, who was the director of the  VA Tech Hume Center. Tim was a research scientist at the Hume Center and was working on applying AI to wireless, both in detecting and processing signals. We recognized this as having large-scale applications in the wireless industry and formed DeepSig in 2016. We are not a service company but rather a software product company with services that help companies who use our products. 


How were you introduced to Blu? How was your experience working with Blu?

We were introduced to Blu through the Center for Innovative Technology (CIT) and met Steven Chen and, ultimately, Pradeep Kaul. Pradeep’s experience at Hughes Network Systems was valuable as he understood the technology.  Steven also had key contacts in Asia which helped us immensely. In 2018, Blu invested $250k in a convertible note through its cyberseed fund (CSF). Since that time, they participated in two follow on rounds for a total of $1.96M. Pradeep serves as a board observer and has been a sounding board for me throughout the years especially negotiating terms with our commercial clients.


How has the company changed since its founding in 2016?

We wanted to fundamentally change the way wireless signals are processed. The way to do this was to replace traditional algorithms with AI because signal processing was getting more complicated, beyond where someone can write algorithms to solve this. In early 2018, we introduced our first product, OmniaSIG, a machine-learning (ML) software application that provides RF situational awareness to a wide range of radio systems. It has evolved since then and detects and identifies signals significantly faster than traditional approaches for real-time reaction to mission critical circumstances. We just announced a partnership with Anritsu, a test equipment vendor that will be selling it for commercial applications. The announcement can be found here.  We are also talking to people in the 5G space to actually build OmniaSIG into 5G base stations to help with interference and spectrum sharing.  


We recently introduced our second product, OmniPHY-5G, DeepSig’s AI-powered solution that elevates the performance and efficiency of Open vRAN 5G networks. This software solution brings the advantages of AI and machine learning to the 5G physical (PHY) layer. It integrates with Intel with Intel® FlexRAN™ Layer 1 (L1) reference software.  We replace a lot of FlexRAN’s signal processing with AI to get better performance. It is the industry’s first 5G carrier-grade neural receiver and redefines how the physical layer in wireless networks is implemented and optimized using Artificial Intelligence/Machine Learning (AI/ML). The announcement can be found here


AI has been the foundation of DeepSig since the beginning and has now taken off. How has this interest in AI impacted the company?

The application of AI in DeepSig’s product was the brainchild of Tim, our CTO. When we first introduced our software products using AI, we experienced resistance from many major companies as they didn’t see any purpose for it. We were introducing technology, especially for the physical layer, before the market was completely ready for it. Now, these same companies are partnering with us and adding the code into their systems. We pioneered AI for wireless and created a whole new category. If you have a really early idea, you need to make sure you have enough runway to survive long enough for the market to accept it. Until then you work on your product and build momentum until the time is right. That time is now for DeepSig.


As a multi-tier CEO, do you have any advice for early stage companies and first-time founders?

If you want to have a company that can grow and scale, you need a plan that includes bringing in the right people at the right time. When we first started, I wore many hats from finance to operations and everything in between. As we grew, we hired key positions that allowed me to focus on the strategy and direction of the company. Otherwise, you’re going to self-limit the growth of the company. Hire key players with experience in the space but who are also trainable, and delegate responsibilities. 


How important is the board of directors with providing strategic advice to DeepSig?

I believe in bringing together a board with outside skills. Scout Ventures and Lockheed Martin are on our board and  bring defense experience. Leawood Capital is run by a former Verizon group president so he has commercial experience. We also brought in Marni Walden, who was the COO of Verizon. As I mentioned earlier, Pradeep Kaul brought an understanding of the technology due to his experience with Hughes Network Systems. Having a board with expertise in both defense and commercial has helped DeepSig introduce and scale these products for both industries. 


What does the future look like for DeepSig?

Looking beyond Defense and 5G, we are exploring other wireless systems such as WiFi and the 6G market. We want to become a dominant AI player in these spaces. Right now, we replace about 10% of the software in a base station but we want to get that up to about 40%. And we're doing this in parallel on the defense side to where instead of using standard 5g technology, we've taken the core technology and put AI on both sides of the communication link. The AI is continually learning around jamming and other impairments. This is generating a lot of excitement in the defense space. We think that technology rolls right to 6G to continue expanding commercial opportunities.

コメント


bottom of page