“The mission of the company is to provide the best that cutting-edge technology has to offer, to the people who need it most.”
John Long, Neuroscience PhD Program alum (entering class of 2005)
John Long is a co-founder and the chief technology officer of Qwake Technologies, a company that builds systems to help firefighters see and navigate through smoke and darkness using advanced thermal imaging, augmented reality, and artificial intelligence. Their products also allow commanders outside of burning buildings to see, in real time, what the first responders inside are seeing. Currently, Long is developing AI-based navigation systems to enable firefighters to trace their paths back out of buildings if they become disoriented. Qwake’s potentially life-saving technology was inspired by Long’s experiences doing neuroscience research.
Long first came to UC Berkeley as an undergraduate, where he double majored in neurobiology and philosophy. He did an undergraduate honors thesis on synaptic plasticity in the lab of Helen Wills Neuroscience Institute (HWNI) member Yang Dan. Long later joined the Berkeley Neuroscience PhD Program, where he worked in the lab of Jose Carmena on brain-machine interfaces, focusing on how the brain learns to use a prosthetic device.
During his postdoc at Rutgers University and New York University, Long developed a multi-camera system to track and correlate animal behavior with neural signals as part of his research on memory and learning. He says he became “seduced” by engineering, and eventually used his expertise in computer vision to co-found Qwake Technologies and develop the company’s technology.
In this Q&A, Long discusses his interests in philosophy, neuroscience, and engineering; his path to entrepreneurship and why he hadn’t envisioned it for himself; and his advice for current graduate students. This Q&A has been edited for length and clarity.
Q: How did you become interested in neuroscience?
A: I first got interested in neuroscience when my science teacher recommended the books of Oliver Sacks to me, most notably The Man Who Mistook His Wife for a Hat and An Anthropologist on Mars.
Strangely enough, I got into philosophy at a young age. I was into comic books, and I went out to visit some family in rural Illinois when I was around 12. The ancient Greek gods and heroes, like Zeus and Achilles, seemed to me like the superheroes of the ancient world. I went to this bookstore looking for a book on the ancient Greek gods and the bookseller responded, ‘Well, that book doesn’t really exist, they were not comic book characters to these people. But the ancient Greek section is over there.’ Not really knowing anything, I ended up buying Plato’s Republic. Over that summer, I read the whole thing. It blew my mind — the idea that this person was writing around 375 B.C. and saying far more interesting things than most of the people around me.
Then in high school, when I really got into science and biology, that’s when my biology teacher, Mrs. Peirce, referred me to Oliver Sacks. It really opened my eyes to the idea that our conscious experience is not only physically determined, but that specific brain areas play a major role in what you and I call cognition. Also, reading those neuropsychology case studies really got me into deconstructing consciousness. It got me into this big idea that our monolithic experience of consciousness is actually a bunch of subprocesses that interact with each other, and that it is possible to deconstruct cognition — to analyze perception, action, and higher order cognitive functions — and that it’s realizable in any physical material.
I didn’t even really know what neuroscience was, what the practice of neuroscience would be. But I knew at that point I [would] study philosophy and neuroscience. I liked those topics. It changed my whole worldview. It was kind of a big deal actually, for me.
Q: You also went to UC Berkeley as an undergrad. Did you do any neuroscience research then?
A: I did. I double majored in neurobiology and philosophy. I worked in the systems neuroscience lab of Yang Dan. I was working on synaptic plasticity with her then graduate student Rob Froemke. I’d read some of the work of Donald Hebb — that whole ‘neurons that fire together, wire together’ idea. At that point, I was really interested in synaptic plasticity as it relates to learning, and I had totally unrealistic fantasies about how one could scale it up experimentally. Yang was a great mentor to me. I did a whole thesis and everything. I was really into research.
Q: What made you want to stay at Berkeley for your PhD?
A: I took about three years off between undergrad and grad school. My senior year of college, I lost a friend who was very dear to me. I needed some time to get my head on straight. During that time, I had great friends in the Bay Area and I was really into mountain climbing. I just had a really great community. And I knew that I still wanted to do systems and behavioral neuroscience, which are, I think, arguably some of the most challenging subfields of neuroscience — doing the experimental animal surgeries, building the electronics, and doing all the analysis. I knew I wanted to be in a place where, when I wasn’t working on my dissertation, I could see the sun and hang out with my friends. So when there was the opportunity to come back to Berkeley, I was really excited to take it. Because I knew that, for one, Berkeley’s an incredible place. I felt very touched that I was invited back. But also, the quality of life is great. I interviewed at a couple other places; I was kind of curious about checking out the East Coast. While I was really impressed by the caliber of the minds out there, they didn’t seem as happy, at least at that point in my life. So that’s what made the decision for me.
Q: What was your experience like in the Berkeley Neuroscience PhD Program?
A: I was in the lab of Jose Carmena in the electrical engineering department. Jose was a new PI at that time, and being one of his first graduate students was a fantastic adventure. Once I started doing my laboratory work, I was down in the animal facility — underground in a room with no windows, working away and all that [laughs]. But the lab I worked in for my day-to-day analysis work was up in Cory Hall for a while. It was in Cory Hall where I got my first glimpse into the cutting edge of computer vision and robotics.
So mainly, my interaction with the rest of the [Helen Wills Neuroscience Institute] was going to colloquia, going to Granlibakken [for the Berkeley Neuroscience] retreat, and going to organized social events like the beer and pizza they’d have out on the lawn of Helen Wills.
Then [there was] all the great admin staff. I particularly remember that Kati Markowitz and Alexis Kurland were really great. When I was there, Bob Knight was [the director of HWNI]. He was always such a cool character, kind of like somewhere between Frank Sinatra and a mentor. He had this cool vibe about him, but he was also a very knowledgeable neuroscientist and wise about the world.
I really liked it. I don’t think you’re challenging yourself enough if getting your PhD doesn’t break you down a little bit at a few points, but it was a great experience. The more time passes between me and my time as a graduate student, the more grateful I am for it.
Q: Tell me a bit about what you did for your thesis project in the Carmena lab.
A: I’ve always been, and still am, very interested in how learning works; what learning is; how do we define the physical systems that learn; and how it’s an interaction between the innate physical properties of the system and the environment in which it’s in. [For instance,] stress is a major concept relating to cognitive function and learning.
For a lot of my PhD, I was working in a brain-machine interface lab where the goal was to have a person who maybe lost a limb, and they get hooked up with a robotic limb, or that has maybe lost a sense, and they’re getting their brain stimulated in a way that mimics the normal amount of stimulation to the brain. I was very interested in that endeavor and it was a major component of my work. But [I] was also looking to use the fact that when you connect a device with the brain, you’re creating a focal point for learning in the brain. You’re inserting this thing into the brain and going, ‘Alright brain, you need to figure out how this thing works’ so you can do what you want. And so the really cool context, and the majority of my work in my PhD, was about understanding how the brain learns to use a neural prosthesis.
For a lot of it, I worked in rodents doing work on the stimulation of the barrel cortex [(a region of the brain that receives input from the whiskers)] — which is basically like their ‘eyes’, or at least one of the more sophisticated senses — and trying to understand how the cellular physiology around the point of electrical stimulation changes as the animal learns to use that input as a cue for when it can get reward.
I also worked on some of the control algorithms that were used on the primate side of things — working with the people there, and understanding how to take the output of the motor cortex to move a cursor on a screen. That is a big part of [what] the lab is very interested in. The brain is like the ocean, it’s just always active, and I worked a lot on statistical methods for trying to understand the difference between normal fluctuations in brain activity versus something that might be worth looking at as it relates to sensory input, motor output, or cognition.
Q: What did you do right after you got your PhD?
A: I moved to New York City. I decided to continue working in rodents, because there was a glaring gap in my work to better understand learning: I had never really worked explicitly on memory systems. I had the opportunity to work in the rodent hippocampus in the behavioral neuroscience laboratory of Dr. György Buzsáki. He was still at Rutgers in New Jersey when I first joined the lab for a postdoc position, but he moved over to the NYU Langone Medical Center about a year after I joined. That was very exciting to me, and I made that move with his lab.
There, I was really interested in expanding what I had been starting to do a lot of in my PhD, which was using camera-based systems to get a more detailed ‘lens’ upon behavior. Most of the work in behavioral neuroscience is either mediated through what are called manipulandums — e.g., the subject pressed this button so many times, or they stuck their nose in this hole so many times — or it’s the laborious task of somebody looking at a video going, ‘Okay, that subject sat down now, they turned left, they turned right.’ I was always very interested in using multi-camera technology and artificial intelligence to speed up that process by automatically labeling behavioral data.
I had the opportunity to do that in the Buzsáki lab. I got pretty much carte blanche from Dr. Buzsáki to build this multi-camera system where the animals are doing a foraging task, or they had a memory component they had to learn every day — they had to learn this new maze for where reward was to be found. I would record everything from when I was teaching the animal the task, through when I would permute the task every day. So, seeing what I would call capital ‘L’ versus lowercase ‘I’ learning. It’s like teaching a child to play piano for the first time, versus teaching a child a new piece of music. And [I was] recording bilaterally from the hippocampus, particularly dorsal CA1 where there are these cells called place cells. That was what I was investigating there — trying to understand how the moment-to-moment dynamics of the brain and behavior relate to memory and planning through space.
Q: So how did you go from that to founding a company?
A: Yeah [laughs], I know, I’ve had a couple acts! Dr. Buzsáki was fantastic [but] I had a couple things that were a little bit, at least for me, discouraging. I would always want to try to be working on the edge of what was going on in neuroscience. And I had a couple instances where, quite frankly, I would try to base some of what I was doing upon recent work in the literature, and sometimes I had a hard time recreating results that I thought I should have been able to recreate. That, coupled with the fact that I was doing very complicated and difficult experiments. The reality [for] anybody doing in vivo physiology in rodents that has a serious behavioral component [is that] only about a quarter of the animals make it all the way through pre-surgical training, surgery, post-surgical rehabilitation, and full behavior. These surgeries were like eight to ten hours, so when only about a fourth of these surgeries were working, and I would get all this data and not be able to recreate some published results — it burnt me out a little bit.
From my work in the brain-machine interface lab, and through all the work I did in high-speed videography and 3D reconstruction, I was becoming quite expert at what is now called computer vision, and in a real-time image processing context. When I started seeing what was happening in augmented reality, I thought, well, the easiest way to get information into the brain is through our eyeballs. I also got a little seduced [by] the satisfaction of engineering. In science, there can be such a long lag between when you do the work and when you know it’s solid. Whereas in engineering, you get a lot of satisfaction along the way, because you build a lot of stuff that just works, or you know when it works, very clearly. So, that led me to reconsider what I was doing. I started getting interested in technology ventures, and started looking for opportunities outside of the standard academic track. That put me on the path I’m on now.
Q: What does your company do?
A: My company, which I co-founded with my partners, is called Qwake Technologies. The mission of the company is to provide the best that cutting-edge technology has to offer, to the people who need it most. For example, our flagship product that we’re building right now is for the fire service and first responders. The problem it addresses is the fact that it’s often not fire that kills you, it’s the smoke associated with it. It creates major visibility problems, not just for the victims of that fire, but also for the fire service members who are doing search and rescue. Every year, there’s several line of duty deaths. Sometimes it’s just a two-story house with low visibility, and because of all the plastics in our homes, those fumes are toxic, and they get disoriented. We forget how quickly you can get disoriented if you can’t see. Tragically, they can end up sucking down their air and suffocating.
We’re using a combination of longwave infrared cameras (because those cameras can see in the dark and through smoke) and augmented reality optics to restore vision to the fire service members. Along the way, we learned that the incident commanders outside are always worried about their people. If I’m an incident commander and you’re a firefighter, as soon as you go into a smoke-filled building, I don’t know where you are anymore. If I want to know what you’re doing, I have to bother you on the radio, and radio traffic is a precious commodity in the fire service. We have things that already exist now: live video streaming of that thermal feed. In a lot of cases where cellular service is available, the incident commander has a tablet application, and they can see what their people are seeing. We have built a whole suite of wireless and cloud-enabled features to improve group visibility, communication, and coordination.
What I’m working on now (and the big thing that keeps me up at night as I’m working on it) is a navigation artificial intelligence. In the interior of buildings, GPS doesn’t work. We have a pretty decent little computer on the firefighters’ heads, and I’m advancing the technology I developed (even during my postdoc, for doing 3D reconstruction and multi-view camera geometry) to basically use the camera on there to reconstruct in real time the trajectory the firefighter took, in a map of the environment in which they’re navigating. So if they ever get disoriented, even with the thermal camera, they can turn on the AI to do what we call ‘breadcrumbs’ — they can look back from where they came and it’ll compute and visualize them the trail back to where they came in.
We’ve done a lot of prototype iterations, really proving out our product at fire training facilities with the gracious fire departments who’ve invited us, that they’re interested in what we’re doing. We’ve got our features really nailed down, we know the firefighters want our product C-Thru, and now we’re doing the hard engineering work to make sure that our system can survive in the pretty extreme environments that firefighters operate in.
Q: When you were in grad school, did you have a sense that you might want to go into industry and in particular, entrepreneurship?
A: I’m not the person to ask for the 10-year plan, for sure. My path has definitely been pretty circuitous. In graduate school and work in general, I figured out who I am and what I like as I was doing the work. Then as interests diverged from the path I was on, I changed direction.
I noticed that I was enjoying the engineering part a lot, and that I was getting kind of burnt out on the animal physiology, to be honest. I also was getting a little bit discouraged by how I was seeing the academic job track. Going from being a postdoc to a PI, to my mind, seemed like going from being an expert-level scientist to being a mediocre manager, and that just didn’t really appeal to me. I love doing the hands-on work. For example, in my company, as much as my CEO and I have disagreements at times, it’s nice that he’s the CEO. He deals with investors, raises capital, and even when it comes to hiring, he’ll do a lot of the initial work. And I’m the CTO, so I get to keep doing the technology part, which is what I’m best at. I like the independence; I always found the appeal of having my own lab. I enjoy running teams of people working on a shared vision, or at least topics of shared interest.
I didn’t grow up in an entrepreneurial family; I grew up in a very working class family. My dad was a ticket agent for the airlines for 33 years, and my mom was a bank teller for a long time. So entrepreneurship, having my own business — it always seemed like something that a different kind of people did. But over time, I was like, why not? It seemed to check the boxes. It’s a really exciting time in technology too, there’s just so much going on. It seemed like there was a broader terrain for my skill set. Actually, given my temperament, I felt like I had more opportunity in entrepreneurial technology than in academic neuroscience.
Q: Do you have any advice for current graduate students?
A: I’m being relatively candid because my switching from academic neuroscience to entrepreneurial work … part of it came out of some critiques I have of modern academia. That’s not a dig on Berkeley or the neuroscientists there. I think the caliber of the education and training one gets in a good graduate school program like Berkeley is indispensable. It’s just so valuable. While I proudly consider myself a neuroscientist, academia wasn’t the right fit for me at that time.
The thing I would tell graduate students is, first of all, realize your adventure is very open and value the time you have in grad school. It’s not just about becoming a neuroscientist, it’s about becoming a scientist and learning the scientific way of thinking, and also really accomplishing something. Berkeley’s hard, they don’t just give you your PhD, you have to do a real body of work. I appreciate how powerful that is, because when it is done no one can take that away from you. You have to earn it at Berkeley.
Graduate students can sometimes get a little despondent. They feel they are only trained to do one specific area of work on one specific topic. The thing I am glad that I did in grad school, and I recommend all grad students do it, is about the third year in, take a piece of paper, divide it down the middle into two columns, and take some time to go through all of your skills and the things you do on a daily basis. On one side, write down the ones that are really only relevant to your particular question and work. In my case, I’d been doing rodent brain surgery, but I’m not a doctor, so I’m not a brain surgeon, and I’m not a vet, so I can’t fix other people’s rats. It really was only something that was relevant to systems neuroscience. On the other side, put your skills that are not just specific to the area. For me, that was computer programming, applied mathematics, and statistics.
When I did that, I always remembered that grad school is still school, and I started allocating a fixed amount of time and resources, and didn’t feel bad about sometimes not working specifically on my project and [instead] being like, ‘I’m going learn this area of math’ or ‘I’m going to really focus on this programming problem’. It was very empowering, and I recommend that to all graduate students because graduate school is not just professional training — it’s bigger than that. And I think Berkeley is a great place to give people a sense of opportunities, or at least the context in which to develop the skills to be successful.
Q: Is there anything else you’d like to mention?
A: Yes, [our company] is growing, so if anyone from Berkeley is interested in this kind of work, [we’ll] always give people from Berkeley a look. Otherwise, I hope they all enjoy their time!
By Rachel Henderson