A Circular Economy is still a distant dream unless technology can deliver smart products that can revolutionize the waste recycling industry with the intelligence of a domain specialist.

Vatsal Kansara, our next pathbreaker, Computer Vision Engineer at Ishitva Robotic Systems, works on AI based Waste Solutions like Smart bins, Air Sorter & Robot that automate the picking, sorting, segregation and recycling of waste materials.

Vatsal talks to Shyam Krishnamurthy  from The Interview Portal about being drawn to Machine Learning due to its ability to learn to do tasks like a human, but with precision and at scale.

For students, understand the vast applications of AI and Robotics in the real world and take up an area that can benefit from these technologies.

Vatsal, tell us about your background?

I was born and brought up in Ahmedabad, Gujarat. My family has a musical background passed on to my father and uncle by my grandfather. My father holds the degree of Visarat in tabla. The main business of my father at that time was a photographer and editor while working part-time as a tabla teacher in school. In my childhood, I grew up around large old cameras, cassettes, and computers. At an early age, I had very good exposure to computer software and various computer-related technologies. Hence, I developed a keen interest in how these electronic things work. I always liked breaking things down and seeing every component of it and then reassembling or upgrading it. During my teenage years the photography industry was on the rise. So my father used to teach me photography o I could help him in case he was overloaded with work. My aunt used to teach me drawing. So, from my early childhood days I had a very good hand in drawing

What did you do for graduation/post graduation?

In school and college days, the practical sessions were my favourite because I could see things in action. I think practical studies helped me a lot in learning during that time. I graduated with computer engineering from Gandhinagar institute of technology, Gandhinagar, While doing graduation I started working as a part-time tutor at my previous institute from where I studied 12th science. I was always interested in physics, especially cosmology and its integration with the field of mathematics. So these interests kept me hooked to teaching.

While in engineering I kept learning new technologies. I used to spend 2-3 hours a day reading blogs and articles about new technologies and trying to implement them. I got to meet 2-3 friends in college who had the same interests as me and we shared lots of knowledge about software, programming and hardware. I started volunteering for the Arduino club of our college and started working on small projects to gain knowledge of the hardware side. In my final year I also coordinated a workshop for new students to share my acquired knowledge. Being a geeky guy, during my leisure time, I started to develop an interest in 3d modeling. While observing films and editing techniques from my childhood, I already knew some basics and extended that knowledge and made some animation videos. During various celebrations at college, I used these skills to produce some videos for celebrations in college. My exposure to 3d modeling and animation drove me to game development , where programming and 3d modeling meets. So basically, I started learning Unity game development which led me to make some basic level 2D and 3D games.

At the end of my graduation I got introduced to machine learning from my frequent blog reading and research. I got to know that this field was in an early stage but it had achieved something as exceptional as image classification and object detection. Programming a computer which can learn to do tasks like a human, isn’t it amazing?! So I started learning machine learning and tried to find the best resources for acquiring knowledge. I came across this course by AndrewNg and I think it has taught me the best methods from basics.

What made you choose such an offbeat, unconventional and uncommon career?

The key influencer and mentor who built my interest in the computer field was my father. He always kept tossing things and ideas related to computers at me. I still remember, during my teenage days, whenever there were any problems with the computer, my father used to give me his computer to repair. He always used to tell me that the internet was the best teacher.

I had a very good interest in reading as well, I used to go to the library near my residence during Sundays. I mostly read science fiction or biographies of people like Steve Jobs, Elon Musk.

I think the turning point in my career was when I came to know about the machine learning field and the course of machine learning by AndrewNg.To some extent, the course had a learning method exactly as i had needed , applying all the knowledge gained practically from the foundations.

How did you plan the steps to get into the career you wanted? Or how did you make a transition to a new career? Tell us about your career path

In the last year of our college, we had a project on CRM (Customer relationship management). I worked on that project (Java based) during an internship at croods consolidates. In that internship, I got to understand company structure, different roles of programmers, and how projects work in big firms while continuing to gain knowledge about programming. At the end of the internship, I had good exposure to client-server applications, project work and standard procedures to create applications based on client’s requirements.

While working on the projects I was also working on my machine learning skills. So, after a year, I decided to advance my career toward this field. I had also continued teaching during all these years and ended up being a math teacher at Sandip Patel institute. As my interest was in a different field, I decided to quit my job as Java developer and started acquiring knowledge about machine learning. For two to three months I worked on my skills and started giving interviews.

How did you get your first break?

Finally, I got a job at Ishitva Robotic Systems as a Trainee Machine learning Engineer. In my starting months, I worked on gaining more knowledge about how these machine learning algorithms work in real life, especially the standard processes 

What were the challenges? How did you address them?

I think the major challenge in my career path was to keep learning and upgrading myself with new information. Other than that, you also need to develop good communication skills if you want to keep yourselves in the limelight. I believe that presentation skills are equally as crucial as your working product.

Where do you work now? Tell us what you do?

I am currently working as a Machine Learning Engineer for Ishitva Robotic Systems. Ishitva robotic system is an organization which provides technical solutions to waste management firms. My work here is in the area of computer vision, which is a subset of machine learning. During my Initial days at work at the company I learned the standard procedures while gaining knowledge about computer vision algorithms.

My first project in the company was to come up with IoT Based low power solutions for smart bins. Some of the challenges that I had to face were power and connectivity. Because these devices need to continuously monitor bins, we can’t consume more power or rely on batteries as these devices are large in numbers, so the overall cost will be very high.

The connectivity should be robust so we get very accurate and real time readings. So, I solved this problem by using custom boards.

The basic working of a smart bin is to get the fill level data in particular bins. This data is used by the waste collector of an area to decide the route on which he/she is gonna take the garbage and at what time. Also the supervisor of that area can verify that all waste management operations are working properly. The smart bin also has a feature to get information about temperature and humidity data in case of any fire emergency.

As I gained knowledge about computer vision algorithms, I started to work on the major product of our company which is the AirSorter and Robot which can sort waste from the conveyor at high speed. My role was to provide an intelligent solution for both systems to gain maximum information about the things on the conveyor.

I am part of the team involved in the computer vision aspect of the system and provide nth level information about the items on the conveyor. We work on training the systems to be accurate and speedy so we can sort the maximum amount of waste to recycle.

My major tasks of the day consist of working on and understanding research papers which have been published by machine learning engineers across the world. I develop machine learning algorithms to achieve state of the art performance of our dataset of millions of images of waste items, while continuing to upgrade the dataset with new information as required and managing it.

The thing I love most about this job is that there is no limit with respect to computer vision technologies and every day there is something new to learn and experiment.

How does your work benefit society?

The Problem of waste management is humungous as India generates nearly 26,000 tonnes of plastic waste every day. Ishitva Robotics is the company which provides technical solutions to waste management companies with its AI powered Robotic System and Air-Sorting machine. By using AI algorithms we try to segregate large amounts of waste for recycling so reuse can be maximized. For example, one Air-Sorting system can sort 5 tons of waste through the conveyor upto nth level of segregation, including giving the insights about brands.

Your advice to students based on your experience?

”Never stop learning in life ” and read lots of books.

Future Plans?

My future plans consist of researching more about these computer vision algorithms and implementing them on new types of hardware.