June 25, 2017

Soodeh Farokhi Makes Robots Smarter with C2RO

Interview by Anjali Bhardwaj


Co-founder and Technology Lead at C2RO, Soodeh Farokhi, holds a PhD in Computer Science from Vienna University of Technology, Austria, as well as a Masters and Bachelor degree in Software Engineering from Shahid Beheshti University, Iran. As a software project manager of two large-scale projects, she has obtained over 6 years of industrial experience in the telecommunication and web service domains.  She has previously done research on cloud computing, big data, and distributed systems.

C2RO (Collaborative Cloud Robotics) is a robot-agnostic software robotics platform that uses real-time stream data processing technologies to provide server-based AI-enabled software solutions for any type of robots. The startup is currently incubated at TandemLaunch.


Soodeh Farokhi, Co-founder and Technology Lead at C2RO


Tell us about C2RO.

C2RO is a Montreal-based startup that aims to augment the perception and collaboration capabilities of robots.  Our platform connects robots and processes their sensor data in a fast, secure and reliable manner. We offer three types of software services: robotic vision, multi-robot collaboration and sensor data analysis.

Ultimately, C2RO gives meaningful intelligence to any robot, even ones with limited sensing or computing ability. The way we do this is through a hybrid approach that places the main part of the robot brain is on our cloud platform, while keeping peripheral computation on the edge. This eliminates the need for very expensive onboard computing resources such as GPU (Graphical Processing Unit) or high-end sensors. Once the brain of the robot connects to our platform (the centralized brain), the robot can use all the technologies that we offer in computer vision and artificial intelligence. The goal is to help robots perceive their world better and become more autonomous.


You say your platform works with any kind of robot. Tell me about that.

Our solution is compatible with ROS (robotics operating system) that is highly utilized in the robotics industry. The analogy is similar to Android-compatible applications for smartphones. In the same way that an Android application is independent from smartphone vendors such as LG or Samsung, our ROS-compatible solutions are agnostic to the robot vendor. Therefore, enterprises with heterogenous robots would benefit from our solution to manage their fleet of robots with a unified cloud-based dashboard.


Did you always know you wanted to build a company? How did you end up here?

My first experience founding anything began with a conversation between some of my colleagues and my supervisor from my time completing my master’s degree. Together we started the first Service-Oriented Architecture (SOA) research-industrial institute in Iran. SOA is a style of software design that is essentially a collection of web services that communicate with each other through a communication protocol over a network. It was an amazing experience building the foundations of this institute from the ground up, building our brand. After only a year we were consulting for national scale government projects and it felt great, however, entrenched in logistics and corporate analysis in a service-type startup, I developed a yearning to return to the world of deep technology. That yearning led me back to academia to complete my PhD in Cloud Computing at the University of Vienna. However knowing that I wanted to eventually apply what I learned to real world problems, I immersed myself in Vienna’s entrepreneurship community. Through numerous workshops and startup camps I equipped myself with the knowledge I would need to build my own deep technology venture.

Nearing the end of my PhD, I started to look into incubators around the world that would help me realize my dream of starting my own venture. I was ready to move anywhere for the right place. I got to know TandemLaunch and its model via a friend of mine, Tara Akhavan, who is now a co-founder and the CTO of IRYStec, a graduated startup from TandemLaunch. In particular, what I really liked about TandemLaunch was their investment on talented entrepreneur whi did not need to have an idea when starting the EiR program, their focus on building deep-technology startups, along with an invaluable business guidance. So I made my mind and moved to Canada to begin my journey with them!


Could you tell me about your research in Cloud Computing, especially Cloud Applications, and how it helped you create C2RO.

My PhD provided me with a multidisciplinary background, a combination of mechanical engineering and robotics, and software engineering, which allowed me to focus on self-adaptive cloud application using control theory, AI and autonomic computing in an experimental setup. My goal was to make cloud applications intelligent enough to auto-allocate the right amount of computing resources – such as virtual machines or CPU and memory – to process incoming requests at real-time with a guaranteed response time as it is needed, while deployed on the cloud infrastructure. This way, a cloud application’s owner would only pay for the required cloud resources to guarantee their application’s performance and not more.

My doctoral research brought down borders between myself and the international academic community. It also helped me recognize my primary interest in a multi-disciplinary, experimental topic, generating within me a passion for utilizing AI technologies to build big data-driven, customer-oriented products that improved lives. Based on my previous experience working for large companies, I was confident that establishing a high-tech start-up while working on Big Data, Cloud, and AI would culminate in my dream job.


How can robotics engineers use C2RO platform and services?

Well, let me tell you a typical story in the robotics industry. John is a robot manufacturer, an expert in robotics and hardware. His goal is to build a robot that provides users with a great experience, but knows little about new technology trends in cloud computing, computer vision, and artificial intelligence, which could greatly enhance the way users interface with his robots. Simultaneously, the needs of his customers are becoming increasingly complex, demanding ever smarter robots. How does John make his robots smarter without building them all over from scratch with novel sensors or additional computational capacity at an enormous cost?

C2RO provides a fast, secure and scalable solution for John with an end-to-end software-only solution with a simple integration process. The complex AI solutions that consumers demand can be integrated into John’s robots by connecting them to our brain in the cloud through by installing a very light SDK (Software Development Kit). A simple monthly subscription to our services replaces the exorbitant cost of overhauling every robot product via an enterprise SaaS platform.


How do you go about improving your platform?

Our cloud-based platform is getting ready for beta release, so request for early access are available via our website. Moreover, having a very strong technical team of experts along with our wide collaboration with research groups from top world universities are giving us a great opportunity to tackle the technical challenges on the way. We are building a C2RO Cloud Robotics Community and are inviting all the experts in cloud robotics, AI, and computer vision to join us to have an impact on the robotics industry.


What do you wish you had known before?

Initially, I thought that finding an innovative idea or technology would be the hardest part of building a company. I figured that the rest would be challenging but that once the ambiguity of not having an exact direction was gone, it would not be so bad. I now know that I was entirely wrong: the hardest part of building a startup is the team. Hiring is shockingly difficult when trying to find the right combination of cultural fit, entrepreneurial sense of adventure and a real commitment to what we want to achieve.

In the startup world, the traditional, passive hiring approach doesn’t work. It is too slow! You have to recruit, tap your network. Also, it’s important not to get caught up with overhead. Forget slow hiring pipelines: just reach out and start a conversation. Also, focus first on the recommendations of your network. They know what you’re trying to do, and so, can help give you knowledgeable references.

Incidentally, that brings me to one of the surprising, positive things that I learned about startups. People want to help founders; it is part of the beauty of this world. They see that you’re passionate about what you’re doing, and if they believe your vision of the future, they want to be part of it, even if it’s just in some small way. Passion is contagious. If you truly believe in what you’re doing, you’ll draw people in to join the adventure with you.


The C2RO team in March 2017.

Leave a Reply