July 5, 2017
Interview by Anjali Bhardwaj
We were lucky to talk to Ashok Prabhu Masilamani, CTO and co-founder of Stratuscent. Ashok has a broad background in various disciplines including startups, IP, electronics, photonics and nano-fabrication technologies.
Located in Montreal, Stratuscent is a startup that is developing the first consumer grade “electronic nose” platform. A TandemLaunch graduate, Stratuscent’s technology is an AI driven chemical sensor that can identify and quantify the ubiquitous chemicals that make up our environment.
Could you briefly tell us about yourself?
I am the founder and CTO of Stratuscent. I have a Ph.D. in Silicon Nanophotonics from University of Alberta which covers a relatively broad mixture of electronics, optics and sensors. Before Stratuscent’s creation, I was the CTO of Applied Nanotools, a semiconductor startup from Edmonton, Alberta. At Applied Nanotools we built X-ray optics and photonics-based chemical sensors. I was leading the chemical sensor development efforts; we were trying to address chemical sensing in the Oil and Gas market – very relevant to Alberta’s energy sector.
While my current interest lies in chemical sensing, I’ve worked in almost all regimes of the electromagnetic spectrum from DC to soft and hard X-rays. I studied at the National University of Singapore and obtained a Master’s degree in RF design and Fiber-Optics. Subsequently I spent some years as a circuit system designer where I designed array multi-processor boards for radar processing and video conferencing systems. With over 20 scientific publications in various fields, you can say I’ve been all over the map. Applied Nanotools connected my expertise in photonics to chemical sensing, positioning me to eventually take on the consumer chemical sensing market with Stratuscent.
How did Stratuscent develop?
There are a number of steps that convert an invention into a product, and those steps are non-trivial, however, they get trivialized by the “innovation ecosystem” that has pervaded many universities and R&D centers. I call what these institutions do “push commercialization” – essentially, inventions are ‘pushed’ out into the market without any clarity towards what problem they address, or if they address a problem at all.
At some point, I learned this lesson the hard way, and came to realize that the better approach is “pull commercialization”. Don’t concern yourself with the technology, instead search for the market’s pain points and then work backwards to see how those pain points can be addressed with the appropriate technology. The market pain point is usually related to a scarcity that needs to be addressed, at an appropriate price point, with a scalable approach. The technology, or invention, must meet all these criteria.
In early 2014, I found the pain point I’d been searching for in the chemical sensing market. We live in an era where our smart devices have developed the ability to sense the world in almost every way that we do (sight, sound and even touch) but still can’t understand chemical context.
The impact of such an understanding would be enormous! Imagine a future where your fridge can inform you when foods have spoiled, or your oven is able to tune its own temperature knobs to perfectly cook your food. A future where previously binary devices like your smoke detector can not only warn you of smoke and fire, but also give you useful information about the quality of the air you’re breathing. All these applications could only be enabled with an understanding of the complex chemical mixtures that make up our world.
There were many reasons such a technology was absent from the consumer market – including, issues of miniaturization and a lack of analysis tools that could interpret the complex signals produced. With my previous lessons learned in mind, I designed a business from the top down, pulling together a business plan, a target market, and then began my search for technologies that would transform this plan into a reality. On the top of my list was the Jet Propulsion Lab at NASA. I tried reaching out to them myself, but equipped with only an idea and no infrastructure or financing to realize it, I didn’t get very far.
Through a serendipitous meeting with a friend, I was informed about TandemLaunch, an incubator that specializes in working with research institutes to spin out technologies: what a great fit! However, after meeting with the TandemLaunch partners I was initially met with some skepticism, which wasn’t particularly surprising. I wanted to approach NASA for a license to a chemistry technology that had only previously seen use in the international space station, a combination of regulatory hurdles and technological risk that was the epitome of a high-risk venture. Luckily, the TandemLaunch model played a hugely beneficial role. One of the key directives of the TandemLaunch incubation process is the heavy incorporation of industry validation in decision making. This meant that the Partners didn’t say no, they simply asked me to validate my proposal with real industry feedback. Better still, they gave me the opportunity to do it through the numerous industry visits to TandemLaunch offices that occur several times a year. It was the extraordinary interest of a multinational IT equipment company worth over $10B that gave the proposal the momentum it needed to get TandemLaunch’s green light. For me, TandemLaunch’s model mirrors the essence of the “pull commercialization” methodology.
You say Stratuscent is developing, “an electronic nose.” Tell us more about that.
We are building a chemical sensor that will enable smart gadgets to smell. Using chemical sensing technology and powerful artificial intelligence algorithms we can identify different chemicals in our environment.
The fundamental sensor technology is the invention we licensed from the Jet Propulsion Laboratory at NASA. Since then our team has seamlessly integrated the power of novel artificial intelligence algorithms, material innovations and robust additive manufacturing technologies to bring ubiquitous chemical sensors.
How is Stratuscent’s e-nose different from other chemical sensors?
In my opinion, the current solutions offered by competitors are fragmented and fall short of addressing the growing demand for chemical sensing. Standard gas sensors like electrochemical, metal oxide (MOX) sensors fall under the lock and key family of detectors – where each sensor is highly sensitive to only one particular gas/chemical of interest. There are other non-real time chemical detection systems that can function like our nose such as, Ion-mass spectrometry and Gas chromatography, however, these systems are immensely difficult to miniaturize and are extremely costly (on the order of $10,000-$100,000). This leaves a large unaddressed market gap for a low power, inexpensive, real-time, miniature, multi-gas detection solution; something that can function like our nose and be at a price point that is feasible for consumer applications.
Incidentally, our technology works much in the same way our nose works. When we smell something, our nose generates a unique combination of signals that form a fingerprint which is processed and memorized, then used to re-identify future smells by our brains. Similarly, our sensor chip is an array of nanocomposite materials that will create a unique digital fingerprint when it is exposed to any chemical – it doesn’t matter if it is a simple molecule – like acetone – or a complex smell – like coffee. This digitized smell is then sent out to our cloud based AI engine which then classifies the chemical, or chemicals, in the mixture.
We are acutely aware of the high-profile failures of e-nose technology in the past. The last decade was filled with big promises in e-nose and not much in the way of delivery. We took this in to account when we worked out our business plan. To realize our platform, Stratuscent is taking advantage of three major technological advances that occurred in the last five years. First, we’ve designed our sensor to specifically take advantage of scalable additive printing techniques that enable volume manufacturing of low cost sensor chips. Second, our computational platform takes advantage of modern advances in cloud-based on-demand deep learning and machine learning algorithms that are available cheaply. Finally, our strategy is to capitalize on the IoT-driven connectivity of appliances and gadgets to enable seamless interconnection between the sensor and our cloud-based AI engine. The optimized architecture from Stratuscent which seamlessly connects all these seemingly separate technological advances plays a vital role as Stratuscent’s unfair advantage.
What was your role at TandemLaunch?
During the incubation period, I primarily lead the business development and sensor hardware development. Although my background was more suitable for hardware development, with mentoring from TandemLaunch’s Partners I was successfully able to lead our business development efforts.
In addition, with mentoring, I started managing the IP in-licensing from universities. As a founder, I was also constantly pitching our business model to LPs and investors that were part of the TandemLaunch family. This helped to prune our pitch and got us ready for graduation. TandemLaunch helped me to shape my role towards a business oriented CTO.
In what ways has TandemLaunch helped accelerate the growth of Stratuscent?
Crucial to any startup is having mentors. Mentoring enables new ventures to solve problems that experts have already gone through. Likewise, TandemLaunch is present from the initial stages of your startup, and has been tracking your progress since those early stages, with that mentoring becomes more efficient because now questions can be answered with context.
This focus on mentorship extends beyond the incubation at TandemLaunch. Anyone familiar with the TandemLaunch model knows that at the later stages of incubation, the venture recruits itself a CEO. I’m a strong proponent of hiring industry veterans for startups that are looking to scale. When you build a startup, scaling is a problem that many entrepreneurs come across which can spell rapid death for that venture. The need to scale pervades your entire business, whether it be manufacturing, talent or sales pipeline. People who don’t have experience with scaling a business may be resistant to the idea of hiring their own boss, because they are typically unaware of the huge number of variables involved. Scaling is also a non-linear process.
Let me explain this non-linearity with a simple example. Imagine a farmer with ten cows. The farmer needs to transport the materials needed to manage and feed them, and transport the milk to a co-op to sell it. Let’s say he wants to expand, so from 10 cows he goes to 20, and so doubles up on everything and sees linear growth. He tries again and this time wanting to go to 40, and that’s when things start to go wrong. When he crunches the numbers, he realizes he can’t go from 20 to 40 because the current vehicle to transport the cows only fits 20 and the next size-up fits 100 at minimum! The size of the transport vehicle is the variable that determines the next profitable scaling point. In a startup, there are tons of non-linear variables like that, and you need to surround yourself with people who understand how to scale. Human tendency is to project numbers linearly unless experience has taught them otherwise. This is why we recruit an experienced CEO, and even an experienced VP engineering. Their previous trials are of huge value to a startup. Nobody can buy you that experience, and it can’t be contracted out to consultants either.
How important do you think incubators are for tech startups?
I don’t have a straight answer to this question. TandemLaunch cannot be classified as an incubator in my opinion. I would call TandemLaunch a “company builder” rather than an incubator.
Typically, incubators provide infrastructure for startups, such as the office space, and administrative support. What most incubators don’t provide is mentorship, acceleration, and funding. I believe there is a strong need for these “company building” incubators to foster the creation of deep tech startups. If appropriately done, what founders can get out of places like TandemLaunch is the ability to handle uncertainty.
If you want to build deep technology companies, you cannot just simply invest in them and hope that it will blossom out. There needs to be a lot of mentorship from people that understand the issues of your company with context. Personally, I think the “company building” model is one that more incubators should follow.
Thank you for your time!