Jetbrain Robotics — Through my eyes
I’m from India. I’m an engineer. We are a billion+ people, producing a million+ engineers every year with a few hundred thousand jobs waiting for them. So if you have a comfortable job as an engineer, you are a rarity already. And being a college dropout with that job, I’ve always had stories to tell.
But this one’s different. We are talking about a tech startup. Particularly, Jetbrain Robotics. I’d say the idea was simple - Build robots (or intelligent machines) to help people. But the execution wasn’t. And so here’s a recap of what transpired over the past 30 months but be warned, this story has no end.
The Birth of an Idea
The bugle call was made on an Indian TV show for startups, in 2017. They asked me if I had to do a startup, what would it be on? Without thinking a lot, I said Robotics. But the more I considered it, the more I realized that it is really what I believed in. Now when the world doesn’t believe you, the best thing to do is put your time & money where your mouth is.
You see startups are experiments. Experiments to see if we could carve out opportunities to create value in society. So here was my experiment to find a niche, identify problems & solve it for the Indian market. And the worst-case in this experiment I imagined, would have been to pivot by targeting a different market instead of India.
I researched how robots around the globe have been helping people, and why India, had minimal of them. The most apparent & obvious answer to the question was that India couldn’t be market to such a premium product.
So I went the other way round. I looked for the most premium robots in the world that were deployed in India. Intuitive Surgical Inc. with its product Da Vinci came out on the top of the list. And to my surprise, it was available in my next-door hospital. The obvious answer just became not so obvious.
Hustling for Answers
My sister is a doctor & and I used to bounce ideas with her. So an idea was to build robots for Healthcare, primarily for the transportation of materials. Although robotic transportation in hospitals was never a new concept, it wasn’t that prominent either (this was the pre-COVID era). I broke my piggy bank (of fixed deposits & PF) and decided to get serious about it. I gave myself enough time to believe that every single penny will be lost. And I’ll have to earn back all of it, again.
So at home, I turn my room into a mini-research rig, with a set time for learning, researching & meeting people. I even ended up designing something myself using Google SketchUp.
Almost everyone that I knew or met, warned me not to start building the product itself before reaching out to the market. Honestly, being an introvert programmer all my life, this was difficult, as it meant I would have to actually go out and talk to people with no guarantee of results.
I wrote to a lot of hospitals with my questions but never got a response or an appointment to answer them. So armed with videos of robots on my laptop, I started walking into hospitals. I used to lie that I was a college student, researching hospitals & advanced use of robotics, to get appointments with the head of the departments. In one instance where that didn’t work, I directly walked into the room of a Housekeeping head & told her that their CEO wanted her to spend a couple of minutes with me.
A lot of my questions involved understanding the frequency of trips, quantity of materials transported, workforce & their costs involved, etc. The exact answer to these was however not known even to the management. Fortunately, one of the CEO’s from a very large multi-specialty hospital invited me to find it out by studying their hospital. They said you might have to sign an NDA for this as a company & not an Individual, and thus with this study, Jetbrain Robotics was born.
Getting Hands Dirty
The next step was to build a prototype. This required a basic understanding of the technology & more importantly, a team that could build it. The later part was painfully pointed out by a manufacturer when I went to him with my robot design in Google SketchUp.
Rohit, our lead mechanical brainy, joined right after his college into an office, which was partially a storeroom. In fact, when his father visited our office, I was so worried that I’d lose the only team member I had. And over the next few days, Somansh, Harsh, Shubham & Lakshmi Narayana joined us. Now we had a team with a diverse skill set including Mechanical, Electronics & Computer Science.
It’s interesting that the first expense that Jetbrain Robotics made was actually on the purchase of a hobbyist robot (Turtlebot 3 Waffle). We reverse-engineered it completely to understand the design, components involved & underlying algorithms.
And in the next few months, we put anything & everything to test. Learning about navigation algorithms, motion control, power requirements became some kind of religion for us. You could find a guy in the dark corner of a room with a manual of ROS (Robot Operating System) even on a Sunday.
A thing about building a product frugally is that it teaches you to respect things that you may often take for granted. For example, take this one — We had a Raspberry Pi. We had a leaky roof. What are the odds that the day Lakshmi decides to keep it running for machine learning training overnight, is the day that it would rain & leak into the Pi? A hundred percent, in our case.
Anyways, In the next few months, We found a fabricator for our robot body who was ready to work with us in iterations with the minimal units we needed. They said they saw in us, the potential similar to a now popular warehouse robotic startup from Gurgaon.
And within the first six months of our startup, we had built a prototype of an autonomous medicine transportation robot running around in the office. We called it AMRO, a short for Autonomous Medicine (or Mobile) Transportation Robot 0. The robot was, however, far from being deployable at a client location.
Not adding the opportunity cost, we had burnt more than 20 lakhs (~ USD 30k) to reach the prototype. Money had run out, we had iterations pending & we definitely needed some mentorship. The hard phase had just arrived.
Hardware is Hard
A hardware project as opposed to software is much more capital intensive. If a robot has 30 components, then you need to have all 30 of them &, not one less for it to be able to work. And for each component, you would need to find a manufacturer, who can provide it to you within the desired cost, lead time & quantity; And very often, technical support for the components as well.
Many of the investors we talked to either didn’t have the patience or the appetite to work with hardware startups like us. They wanted to see customers before they invested, while customers wanted a fully working product on day one. So we started looking for accelerators that understand hardware, to help us.
YC & HAX were the first things that came to mind obviously. We knew YC doesn’t prefer a single founder startup, plus we hardly had market traction apart from a large hospital that would consider paying if we complete the product. We nevertheless did apply. This is when a new accelerator we hadn’t heard of before, sent us an invite -Brinc.
We applied to Brinc’s accelerator program, got through the ramp-up, and got selected into a 3-month long acceleration program, which included mentorship in Hong Kong. I still remember showing everyone the mail of getting selected. My hands were shaking with excitement.
The announcement of getting selected into a prestigious accelerator was an awesome experience. For the first time, our logo appeared somewhere else than our walls. We felt validated after months of hard work. Everyone in the team saw it as the beginning of a new chapter — Paving way for our first official visit to a bar :)
Brinc’s accelerator program was jam-packed with expert mentorship on various fronts. We got to engage directly with startup veterans, who had immense knowledge in their respective fields. We were introduced to various investors & startup programs available in HK, like the HKSTP or Cyberport, two of Brinc’s partners. They are government-backed organizations that provide great resources to startups. Things were happening so fast, that I often used to tell myself that we should be back here for more, later.
We also visited manufacturing & testing facilities in China all as a part of the program. This gave us a stark insight into what we might need in the near future. An experience like this of seeing how things at scale are built definitely helps to grow your outlook. It’s during this visit that we learned about how startups utilized rapid prototyping facilities to go into the market faster. One thing I knew was that we needed to find a way to iterate faster.
Overall, the program helped us improve our plans for the future, especially in terms of cash flows, project management, knowledge about supply chain & manufacturing. And the program ended with us pitching at the Hong Kong Stock Exchange amidst 300+ investors, tech enthusiasts & digital media houses.
As part of this program, Artesian Capital also invested in us with a commitment to being part of the next raise. Unfortunately, it took us more than 4 months to get this money into India due to local legislation, thus hurting our plans for the year immensely. The ordeal of dealing with the banks & compliance issues was a nightmare. It’s at this point I realized that money movements globally were so painful. Anyways, those learnings are probably for another blog.
Minimum Viable Product (MVP) & Modularity
Back in India, Aadi, Arijit & Pawan joined us. And over the next few months, we put our heads back into the ground & started incorporating the feedbacks wherever we got it from. We iterated over our design drastically to fix the problems observed by people we pitched to.
These months carried a lot of uncertainty, as we were dealing with much larger real-world problems. Everyone in the team required to up their ante. And we kept inviting investors, clients, or people in general to our office for feedback. It’s during this time that we realized that making something for yourself is much easier than making it for others.
Something to cheer us up in between was an award by AICRA (For best use of AI in Healthcare), taking a larger office & turning the older one into a conference room, better doorways (cause our robot simply couldn’t pass through a door in the old one), and a little drone that had become one of the favorite pastimes.
We started talking to hospitals even more, but now from a delivery perspective. They asked us questions in terms of warranty, operations & training which made us engage with better manufacturers & suppliers. The only good part during these times seemed to be getting invited to free lunches by the suppliers.
We also observed that many people connected with the idea of robots being humanlike. Although initially, most of us were opposed to this thought as we believed that unless perfected, human robots are more gimmicky than useful. So we decided to build a robot that would be human-like but really useful — ARYA, the patient care assistant. We built a speech-based chatbot into it, so it could answer questions, guide, serve & monitor patients.
We started taking our products to multiple events and simultaneously made iterations over our designs. We spent a lot of time perfecting our algorithms for motion, navigation & maneuverability. The design iterations with ARYA, helped us split the use case & the core navigation as a separate unit — We called it Core 0. We had thus, just brought in a ton of modularity into our designs.
Pilots, Sales & Partnerships
We had just begun to pitch to hospitals, nationwide. I started taking frequent client visits and honestly started appreciating the grit of salespeople. We started putting out a lot of material on social media to grab some eyeballs. I also started speaking out more publicly about our products.
At our accelerator, a new batch of startups had just joined & I often found myself sharing our experiences with them. Co Incidentally, right at a time when I was to pitch to a multi-specialty hospital in Kerala, our accelerator announced an Indian division in Kerala as well. With our early backers Brinc around in the city, we were now considering to start an office in the south as well. Our pitch deck & plans had just started growing especially (and understandably) since most clients preferred robots for lease than an outright purchase.
Our main focus though was on a hospital in Gurugram which invited us to pilot our robots at one of their locations. Although we were stuck with weird issues of trying to make a Chinese elevator about which even the supplier had no clue about, talk to our robot.
We had run short of money by now & looking to raise another round of investments. And fortunately, investors now seemed more interested, as we had an MVP and were looking to raise more for sales than R&D. A major accelerator & two angels had just shown interest to be part of our next raise, and we were about to reach back to our earlier Investors.
It’s during this period that Val, Founder & Chairman of Solaris Disinfection Inc, a leading disinfection robotics company from Canada reached out to us after seeing our navigation platform over LinkedIn. Although to execute this international order itself, felt like an achievement for us. Little did we know that planes could refuse your shipment after your courier company takes it all the way in cause your Lithium-Ion battery is apparently oversized. Anyway, so Solaris bought a Core 0 from us to validate our technology, invited us to Canada for further orders & thus began a wonderful partnership.
Somewhere in the mid-December last year, parallelly to this, when I was in Canada, NASSCOM set us up in a call with KONE Elevators. Here was one of the top elevators companies in the world, sharing with us their fabulous solution, which happened to be exactly what we were looking for. I was so excited, that I ignored the time of the call & attended the same at 4 AM. The call went well, and today, we are India’s first robotics company to have robots that can talk to elevators & transport medicines across the floors of a hospital, autonomously.
So here you are, all caught up, on us.
And this brings us to somewhat in the present…
COVID & Acquisition by Solaris
Since COVID, Solaris has been working in an overdrive mode to deliver their healthcare partners with the LytBots — a fast, effective tool to help control the environmental spread of dangerous pathogens and keep patients safer.
Solaris & Jetbrain have been working together since the beginning of this year. Although we started with the discussions of automating the LytBot initially, we realized early on that we had a lot more synergies than what we had planned to tap into. Our visions overlapped as well. So a partnership of some form was always in consideration.
As Adam, the CEO at Solaris puts it — “COVID has accelerated robotics deployment by five years”, I believe it has acted as a catalyst in our considerations too. As partners we see this as an opportunity & responsibility, to utilize our potential to the fullest and enable people, especially in this hour of crisis.
In the strategic sense, while Jetbrain, with its core technologies & resources, has the necessary technological muscle, Solaris, with its growing market position and extensive distribution network, becomes well-positioned to help Jetbrain further develop and deploy our technologies.
So now, when my close ones ask me why I gave up driving the car called Jetbrain Robotics, I just tell them this — I’m still in the driving seat, it’s just a bigger car with more passengers now!