Alabhya is my roommate. He came to me a month ago speaking about breast cancer detection idea that if you want to submit for health innovation challenge. I did not give it much importance because he did not have any idea about the technology that would enable it. I felt like he is just another wannabe guy who wants to do something but does not have the skills necessary to make it work, but he had a strong case because his mother had the condition he was strong to build the detection for. I was so much involved with my internship and course work that I did not have much time outside of them. In addition to this, I was involved in Alaska Airlines environmental innovation challenge, NSF Icorps customer discovery program, and HPAIR. So, my idea was to just help Alabhya prepare a list of materials for the prototype funding and let him take care of the rest.
After we received prototype funding, I felt a sunken cost fallacy where I have this opportunity to work on this and the resources to do so, and leaving the project at this time felt like leaving some of the work that I did and the opportunity on the table that I would probably regret later. I also got in touch with Yana, a PhD student at the Electrical Engineering dept. was working on short distance ultrasound technology for robot assisted surgery. After getting to know that it can be used for detecting lumps in breast cancer as well, I felt more confidence in the technology and that I could build it.
Even though we received prototype funding a month before the competition, the real work started three days before the competition. Right after submitting my machine learning for cybersecurity assignment two days late because of a problem with the provided data set, I was very frustrated as I had a lot of work to do. I felt that giving less importance to the health innovation challenge would be the only way I could get out of the situation that I was in without stressing too much. But somewhere in my mind, I wanted to do this as the impact of this project if I can pull off a prototype would be much greater than just doing an assignment for the course.
We ordered all the items from adaFruit and Amazon three weeks prior to the competition. It consisted mostly of general-purpose electronics and development boards focusing on building technology using ultrasound. But further research made me aware that pulling off ultrasound was really hard with the limited time that I had, mainly because the sensors needed higher frequencies than the readily available hobby electronics. I started searching for other technologies and the resistance based technology felt like achievable in the limited time frame. There were no hobby sensors available in the market for resistance arrays. Moreover, all the senses that I found were built on rigid backplate which made it less useful for our use case.
Three days before the competition, I started creating 3D renders for the prototype. Rishabh- one of my teammate for this project as well as my academic project for machine learning for cybersecurity course- was in India visiting his family. He helped me so much with the mini project one checkpoint report that might team consisting of me and him had to submit on that day. This helped me immensely to focus on the task at hand. We had not finalized on the prototype and created renders with three different technologies. One of them focused on ultrasound as per the suggestion by Yana Phd, another one based on a contact gel-based technology from MIT, and the third one based on resistance.
Two days before the competition, me and Alabhya went to the maker space in University of Washington to create the prototype at 7 PM. Alabhya was very focused on the pitch while I was working on the prototype. Initially, I made a very crude square 3 x 3 sensor on a piece of silicon material that Alabhya got from his maker space to test if the velostat works. This was done using a multimeter .
Since I got to know that it would work, I spent the next day working on the code for the sensor array using ESP32. I had to prioritize between 3D printing a case for the electronics and coding since there was no one to help me with the tech. I prioritized coding and had this idea of using a pre-existing container for enclosure. I searched for shampoo bottles at home, but none of them had a sturdy mechanism that would allow me to add the sensor. While searching for a container with re-attachable lid on both sides, I remembered that fleshlight had such a mechanism, thanks to the ads. I immediately ordered one on Amazon with a one-day delivery from 7 AM to 11 AM.
The idea was to map the 3 x 3 sensor array into a 3 x 3 matrix shown on a TFT screen. By evening 7 PM, I started testing the sensor array using the code that I wrote, and it worked well. At night, we went to the maker space. I wanted to solder the headers onto the ESP 32. At 12 o’clock, when the maker space closes, I soldered six headers into the microcontroller. I begged the maker space staff (students) to give me two more minutes in order to make sure that the solder joints are not cold joints which they politely refused.
I came home and worked more on the code until 4 AM in the morning under the influence of the McDonald’s burger and high sugar milkshake. I wanted to sleep at least four hours to maintain my cognitive ability, but I could only sleep three hours. From 7 AM to 9 AM, I was working with the old sensor which did not show any promise. Soon, the fleshlight arrived. My brain started to zone out from all the abuse it took the day before. Unflinched, I started playing Malayalam rap songs from Fejo and Thirumali in my Sony 1000XM3 noise cancelling headphones. I took the outer covering and started attaching the sensor array made of aluminum foil, insulation tape and velostat material. The whole process made me feel like living the scene in which Richard Henricks wore the same earphone for a night listening to metallic music and invented the middle out algorithm in the Silicon Valley series.
Thankfully, Sreejith was at the home. It helped me by making some cereal in the morning that made me go through the day. Additionally, he communicated with Alabhya and carried some of the printouts and monitor to the venue since I was busy with building the prototype. Also, I reached out to my manager at Activision to ask her permission to skip a meeting. She was understanding and allowed me to go to the competition.
I connected the wires from the microcontroller to the sensor and thank God, everything was working properly. The sensor was very flimsy, and it kept on breaking till 2:30 PM. I was very frustrated at this point because the competition had started at 12 PM and the teams were pitching their ideas. I have only one hour to get to the stall before the judging starts. At this point, the device state had wires. In the next half an hour, I finalized the code, uploaded it and started attaching the battery. The battery was huge, so it had a possibility of bending if I attached it to the curved surface of the prototype. I did not have much time to think, so I put my Bank of America credit card in between the battery and the curved surface while putting a foam sheet on top of it to prevent the battery from getting too much pressure. I wrapped insulation tape around the battery to conceal it. This was connected to the microcontroller, and viola, the portable handheld breast cancer screening device prototype was complete.
I started feeling a lot of confidence on the prototype that I made as it had very little chance of breaking due to the concealing that I have done in the last moment. This concealing had a lot of importance as the venue was bustling with a lot of people and if I have not done that, it would have definitely broken in half an hour. I took a bath and then to the venue in a lyft cab.
The venue was bustling with a lot of people. The judging seemed like voting based with at least 20 judges and as soon as I reached there, I started explaining the prototype. This went on for three hours. The judges were asking a lot of questions which I did not have a clear-cut answer to, but I was surprised that my pitch got much better after speaking to few attendees. I had faith in the sensor technology and my credentials as a machine learning engineer that I pushed the boundary of our product by explaining the future directions that we could take as well, and while doing that, I started to feel the potential of the solution that I am proposing. The judges were proposing a lot of good ideas that we could incorporate in our solution.
Alabhya was also taking on a lot of judges head on, and during that time, I felt like Ram Charan and Rama Rao Jr from RRR in the scene where they are in the jungle and fighting off the British very close to the climax of the movie.
By 6 PM, when they started announcing awards, I was peacefully stuffing all the savoury food available at the venue as I have not eaten any healthy food from the last night. I was not cared about the judging any more, as had attended more than 20 competitions like this, and my priority was to feed my body so that I do not succumb due to the insane stress that I had to take the past few days.
Then the awards rolled in. Award for the best medical device goes to “fight for 98”. I felt indifference, it is at this point my sleeplessness was getting to me. Me and my team went to the stage to the picture, came back, I refilled my plate and started eating again. In the next award rolled in. The Spark award goes to “ fight for 98”. At this point, my mouth was full and I was chewing on the way to the stage. This situation was very similar to Alaska Airlines environmental innovation award ceremony where I gulped down a glass of wine from the side stand on the way to the stage to swallow the food that I had in my mouth.
The whole experience was rewarding and gave me a sense of calm, knowing that the problem that I pursued as a potential to impact the lives of so many people especially women who are struggling with this issue. I am also happy that I got to be a source of inspiration for all the people who saw my prototype and thought about this issue. I hope I can work on it more to create a robust system that can be productized.
I thank everyone who helped me in this process and I hope I get to do more interesting work like this.