Amir Rubin co-founded Paracosm in 2013 with a mission to “3Dify the world.”
Since then he’s led the company through a total conversion from fun, visionary consumer product to critical B2B enterprise solution. What started as an idea to 3D map the world for experiential AR turned into a technology and approach that saves enterprise customers weeks of work every day.
Ledge sits down with Amir to see what it was like emotionally to transition from a visionary 3D world mapper to getting measurements inside buildings for industrial purposes and how to grow in the CEO function when you come from an engineering background.
Amir graduated UF Computer Engineering (ECE) in 2003, at which time he co-founded Prioria Robotics during his senior year. At Prioria, he helped develop vision-based navigation systems for small UAVs. Amir’s work in 3D imaging continued in 2010, when he developed a device using 3D cameras to weigh cows for which he was awarded a patent.
Amir was the first employee at Gainesville based Shadow Health, which develops 3D-immersive healthcare training simulations. In 2013, Amir co-founded Paracosm, which develops 3D-mapping devices to digitize reality.
In 2017, Paracosm was acquired by Occipital Inc, an industry leading spatial computing company.
Ledge: Hey, Amir. It’s great to have you on. Thanks for joining us.
Amir: Thanks for having me, Ledge. I’m excited to be here.
Ledge: Can you give your two, three minute background story, yourself and your work, how you got where you are?
Amir: Sure. My name is Amir Rubin and I am co-founder and President of the Paracosm division of Occipital. We were acquired by Occipital a year-and-a-half or two years ago, somewhere about there.
We were founded in 2013 with the mission 3D-ify the world – is what we like to call it. My background is as a computer engineer. I graduated from University of Florida, and I started my first company during my senior year, actually. I started a UAV company with friends and we decided it would be really cool to build a vision-guided UAV. It was a major learning experience, but we were were a bit ahead of our time. This was back in 2003/2004.
We built some cool prototypes, learned a lot. The company lasted quite a while too, my first startup. It just finally kicked the bucket a few months ago. It had a good like 15-year run.
After working on vision-based navigation for drones, I left my own startup in 2009/2010 and helped found another company in Gainesville with other University of Florida researchers that was doing 3D nursing and healthcare simulations. Then in 2013 I had a dream one night where I was like, oh, wait. What if we were able to combine computer vision techniques to create 3D maps of the whole world? And I founded Paracosm to make that happen.
Ledge: Off-mike, you and I were talking about how that vision of 3D-ifying everything morphed from whimsical consumer thought process to B2B enterprise.
I’d love to take that path, both on the concept and the technology trail that got you there.
Amir: Sure. The concept of Paracosm was, how can I 3D-map really fun places to let people visit and explore new environments?
For example, my wife and I – then girlfriend, now wife – were backpacking in Europe and I was like, “How cool would it be to 3D map this castle so all of our friends can visit here?” Kind of the idea of Paracosm when we first started was this whimsical concept to be able to turn all these fun places in the world into a video game level.
Then we were like, well, what if we also turned your home into a 3D model of your home, a 3D video game of your home. You can let your home robot navigate your house really well, and it could let your augmented reality headset play you really immersive video games in your couch. By mapping a castle, we could have people virtually visit the castle or we can do all sorts of augmented reality experiences. Historical visualizations of what life was like looking through a headset.
We had all sorts of very ambitious ideas that we thought 3D environments would enable. But, similar to my first UAV company – this was in 2013 – we were way too early on the robotics, AI and augmented reality front to really get the traction we were looking for.
We did have a lot of early partnerships. We partnered with Google and helped them develop what became the Project Tango smartphone. We contributed a lot of code to the 3D mapping engine for that.
We ended up partnering with iRobot on some of their early Roomba initiatives for mapping, back in 2013/2014. We were their first investment. But ultimately, the overall market was just so early and what ended up happening is we got picked up by some 3D scanning blogs – some of our early work that we did with Tango and iRobot and stuff – and geospatial surveyors and mechanical engineering contractors and construction VDC, Virtual Design Construction managers started contacting us. That’s when we realized, maybe our tech can be applied to real world problems. We ended up pivoting to that use case.
Ledge: I need to ask, was that ever un-whimsically disappointing? For something to become less about the passionate, make-you-smile kind of stuff to real world, B2B application? What was the emotional trail there?
Amir: When we make the transition as a team, it was a very hard emotional transition. So, yeah, it was hard. Exactly as you said. We were saying, hey, the whimsy world of 3D is for a robot to serve me my afternoon tea, or to have dinosaurs jumping around my living room. We’re replacing that with getting dimensions of conduit and HVAC systems and buildings.
It is an emotional transition but the two things I would say that changed were, first, we decided we don’t have to become boring. We’re going to bring some of our whimsy nature to this industry. So we kept our logo, we kept our tagline, and we still try to be a little bit silly in this industry, which helps us stand out.
People see our funky logo, which is… Our company name is Paracosm which means an imaginary world. Paranormal Cosmos is an imaginary world. Our logo is a parakeet cosmonaut. A space parakeet. If you look online you can see our fun little logo. We try to bring some of that energy.
The second element is, people started using our tools for 3D mapping real world projects. Our customers were telling us how it’s transforming their businesses, their projects. How they’re really starting to rely on our system, and how when they encounter a bug or when the system crashes we get panicked calls. That started replacing this emotional stock of going from a whimsy, fun consumer to what’s consider a kind of stodgy enterprise.
It started transforming to this relation like, wait, we’re really helping people do their jobs. People are relying on us to do heavy lifting, whether it’s surveying a forest, a construction site, a factory environment or an office building survey. We’re saving people weeks of work with our new technology, and we’re letting them do things they haven’t done before. We hear from all of our customers.
Once we started shipping and having fanatical users, then now it’s really exciting again. It adds a new layer of meaning.
Ledge: The reach of a business is so much more dramatic than if I’m mapping my house. There’s hundreds of people that are touching install, so I could see how that could start to develop on the relationship front as well.
What’s the technical path from, oh, hey, we thought we were going to be consumer and now we’re B2B enterprise. There’s got to be a whole bunch of speed bumps, learnings and other opportunities in there.
Amir: Yeah. On the technical front, on the consumer side we were using the equivalent of, for example, Microsoft Kinect to do our 3D imaging. It’s an off-the-shelf hardware that we clipped onto a tablet or a phone. So the tech stack effectively was a lightweight. I think we were running our stack on Android or tablets and Microsoft Surface tablets on Windows. Most of the what the app did was a quick, hacky kind of visualizations. It was uploading data, AWS Cloud, and we were processing, doing all of our heavy lifting on EC2 instances and TPU.
As we moved to enterprise solutions, we started having customers not being able to upload data to a third party cloud. We have to keep everything local, and we need to start doing a lot more processing load on our 3D mapping device.
So, we had to build custom hardware device with industrial grade laser sensors. We used LiDARs, the Velodyne LiDARs from self-driving car industry. We interface it to machine vision grade cameras. We have an EC Intel NIC embedded computer in there and we do all of the processing on a high performance C++ stack that runs entirely on our device.
The big difference is, we had to replace our cool cloud infrastructure with a totally self-contained device, but we still find some pretty cool uses from our web knowhow and back in the day. We manage our customers as a fleet of kind of IoT devices, almost. What’s actually happening is, each one of our devices is running… We use resin.io and we’re a proud customer of theirs. We run a Dockerized app.
Our app is a completely contained Docker container running a pretty much version of Linux, but whenever we want to send out a software update we send out a new image to all of our users.
Ledge: You’re a CEO, founder that grew out of the engineering function. We talk to a lot of CEOs that come out of sort of sales and marketing. I wonder, what’s your experience been moving through engineering to having to run a company?
I think founders, particularly on the technical side, maybe struggle with that a little bit. So, do I want to get into business, per se or do I want to be an engineer forever?
How do you do that? How do you toe that line?
Amir: I think it really depends on the nature of the business. Sometimes the answer is you can’t. If the nature of the business is that you really are passionate and uniquely positioned to be the technical innovator, the CTO effectively, then perhaps the right move is to be the CTO and take ownership of the technical leadership and product vision.
Maybe some people, you can find a better CTO and you can take over the business. But in order to make the transition, as the company evolves through the stages, you have to evolve within. So do your co-founders and your senior leadership.
A founding team or a leadership team that worked great for a five-person company might start to show cracks in a 20-person company. That same team might start to crack in a 80 or 100-person team. In startups, these kind of team transitions can happen, as we all know, very quickly. Once you get an infusion of, maybe it’s venture money or your product is starting to do really well and all of a sudden you have to start hiring like crazy to keep up with all the demands.
It really is a different skillset at every stage of the game. My advice to people is just like, what are you really passionate about? What are you good at?
It’s unlikely that you’re going to keep being able to do everything. As the team grows, leadership might mean being really good at delegating, executive strategy and execution. Maybe you’re more sales oriented and you want to run the sales team. Maybe, like I said, if you’re more technical oriented you should be the CTO. Or maybe you’re a really great product CEO and you hire out a VP of Sales and a VP of Engineering as CTO. Maybe it turns out you’re a really great people leader, and you are the glue that holds an executive team together and you just hire really great executives.
It’s always a question of what you’re passionate about and what the company needs for its current stage and future stage of growth, really.
Ledge: In your own story, what were you and what weren’t you in each of those paradigms, and how did you come to find that out?
Amir: For me, I… It’s a good self-reflecting question. I would say one of our strengths as a team has always been we’ve always been able to have a really good product vision and be able to build intuition of what the product needs to do, talking to customers.
We’ve always been good at building relationships with customers and getting our engineering efforts to crank out something useful. That’s always been a good strength.
Me personally, it’s probably more like… I don’t have any, for example, background in sales or marketing. Being able to hire people to run sales and run marketing operations has always been really helpful. Then as you start taking investor dollars, investor relations – which again I don’t… The investors speak a different language than us technical minded, engineering background people. They speak the language of finance and they speak the language of board members. If you’re not able to or not interested in learning that language, it might be a good idea to hire a CFO or to work with other people around that.
It’s been an interesting learning journey, for sure.
Ledge: It always is. Alright. Last question. What’s your favorite, funniest, most excellent failure, phoenix rising from the ashes to success, story along that journey?
Amir: Well, there’s many but it was when we were making the transition to jump onto a heavy duty industrial 3D mapping system.
I took our old system and I went onsite with a construction company who was really excited to be able to 3D map their construction sites. I took our original system built around a Microsoft Kinect type camera, and we tried… They were really excited. They had paid a few thousand bucks for this pilot project, and I just could not get a result because the cameras, for example, don’t work outdoors and the construction site doesn’t have windows. The windows aren’t installed. The walls aren’t built. The project was a total disaster.
So we were like, oh, well we need to rebuild this using these new LiDARs, and the LiDARs were really expensive. They’re 8 or $10,000. We were like, how are we going to afford… At this time we’re running out of money to build prototypes. How are we going to possibly afford this?
I ended up getting in touch with a survey company that was so interested in what we were doing, they shipped us a $10,000 LiDAR unit in the mail and said, “Hey, make this work. We want it that badly.”
We bolted it with wires to a laptop and camera, and it was a total mess of a prototype and we posted a picture of it somewhere. Then a company in Japan ended up seeing it and buying it from us. So we sold them this system built with another company’s LiDAR in it, and they ended up loving it and they’re still one of our closest customers and partners. They’ve been with us since the really embarrassing prototype days.
It’s the little things you learn along the way.
Ledge: That’s why I asked that question, because I think everybody forgets the V.1 days where it’s just this embarrassing pile of garbage. Software, hardware, it doesn’t matter. You make it happen. I think that’s the stuff that we forget about as founders who are just trying to scrape it together and become the proverbial overnight success. That it’s not overnight. It’s over many, many nights.
Amir: Yes. Many, many. We’re now part of the Occipital family but we still run our group as a product group and we’ve been doing it going on six years now.
Ledge: Fantastic story. I love it.
Amir, thanks for joining us. Really cool to have you on here.
Amir: Yeah. Thanks for having me, Ledge.