[Techie Tuesday] This engineer went from healthcare and political psychology to mini ICUs for coronavirus patients
In this week’s Techie Tuesday, we feature Aardra Kannan Ambili, the Co-founder and CTO of AI startup RIoT. The engineer taps AI for building specific solutions, from algorithms for political psychologists to AI systems that monitor respiration.
Tuesday May 19, 2020,
12 min Read
Science and technology were a part of Aardra Kannan Ambili’s life since she was a child.
Her parents, both engineers, started Mediatronics, a software company in Trivandrum, when she was five and her younger brother was just born.
Her love for technology continued to grow along with her, and today Aardra is the Co-founder of RIoT, a startup that leverages artificial intelligence (AI) to build highly accurate respiration monitoring systems. The startup has recently built a non-contact, wi-fi enabled, affordable respiration rate monitor (for coronavirus patients) that can run as a mini ICU unit.
Aardra, 31, has been building products that solve varied problems - wearable devices for women's safety, a device to monitor an epilepsy patient’s attacks, and even algorithms that were used by political psychologists.
Her way of working is simple, and focuses on how a particular problem is approached and solved. It is something she looks at even today when she hires techies.
“I look at how they think about problems and how passionate they are. I am always looking for people excited to solve problems. That comes through in the first meeting itself. I need your energy and positive mindset,” Aardra says.
The early years
Aardra’s parents instilled a love for science and technology early on, but she had many dreams growing up. She wanted to become an air force pilot, a lawyer for environmental activities, and an AI researcher.
School was easy for the academically bright student. By the time she reached Class 5, AI was a “distant and abstract thing”, and that “abstractness attracted” Aardra. Her first tryst with programming was in Class 7, with C and C++. It wasn’t something that she immediately picked up, but she found it “interesting”.
Early on, Aardra also realised that if she continued to stay in Kerala, she would be expected to work at her parent’s firm. And, she wanted to do something “different and on her own”.
“I saw that despite the fact that both my parents were founders and engineers, my father was more respected. I would keep wondering why. The gender bias was something I never understood or appreciated,” Aardra says.
A difficult engineering journey
After schooling, Aardra did her engineering from a private university in Trivandrum.
“I don’t talk much about those four years because I didn’t want to do engineering. Despite having an affiliation for science and technology, engineering didn’t have appeal the way it was taught. But, I wasn’t sure what I wanted to do. I had asked my parents to give me some time to figure out what I wanted to do, but a year’s break was a big no,” she says.
“Every year I wanted to quit. I would keep asking if I could drop out because it wasn’t working out instead of wasting their money,” Aardra recollects.
But she nevertheless finished her engineering in 2011 and got placed at Wipro. She took the next eight to nine months focusing on what she really wanted to do. Her parents kept suggesting that she join their company, but Aardra was sure she wanted to do something that was different.
“My parents were achievers and I didn’t know what my calling was. That was one of my darkest times. Even now, I value myself based on the work I do; I need to say I am someone doing something worthwhile,” Aardra says.
Focusing on AI
During this time, she remembered her childhood dream of becoming an AI researcher. She started looking at different master’s programmes and realised that AI was an upcoming sector.
She applied for a programme in AI at University of Georgia at the end of 2011. Coincidently, around the same time IBM had released Watson and a computer could answer Natural Language questions for the first time.
In the next few months, the Mars Rover landed on the red planet. “All this happened with AI and soon there was a growing interest. But, machine learning and NLP hadn’t picked up yet. ”
Aardra says she learnt a lot more after going to the US in 2012. “The way they promote you to think and discuss multiple approaches helps you connect things and find solutions. Apart from encouragement by professors, I could read about different scientific approaches,” she adds.
Soon, that she began researching problem solving with AI.
From healthcare to political psychology
“I realised technology doesn’t need to be meaningless or just academic; it is about what we can do with it,” Aardra says.
She starting studying the use of genetic algorithms to understand how a drug is absorbed in a human body and what variables work.
This was Aardra’s first time working with AI and health, and she realised there could be a correlation between AI and psychology. Especially in the field of political psychology, where “they interpret what we write to understand the kind of people we are and how we think”.
A few professors and researchers were working on this, but they were doing it manually. They would work on hypotheses like ‘why does Obama think broadly on multiple issues’.
“This particular construct is called ‘expectative complexity’, which helps you understand different writings of different people. It is used to understand which political leader will go to war, based on speeches given by him or her last week. How some individuals have a higher propensity to violence, so that interventions are scheduled. But there was a lot of manual work. It needed a lot of research interns and assistants to score, and was error-prone and time-consuming,” Aardra says.
She decided to leverage natural language processing, and defined an algorithm that could score the text using machine learning and NLP techniques, with good accuracy.
Around 2014, Cambridge reached out and said they were planning to continue the work . Aardra says it was an obscure construct that nobody had heard of but it worked and her work got published.
“For me, it was a big kick. As a woman it matters more because whether you do it or not, external validation for women is tougher,” Aardra says.
Joining a Bengaluru startup
In 2014, Aardra came back to India to join a Bengaluru startup founded by IIT-Bombay alumni.
The death of her dog propelled her return despite doing well in the US. “It struck me that I couldn’t see him, and I was terrified that something would happen to my parents and I wouldn’t be around them. I wanted to be closer to my family and moved to Bengaluru,” she says.
The startup scene in Bengaluru was growing and Aardra was excited by the pace of the work. The startup shut shop, but it pushed her in a new direction: automating the chat conversation. So, instead of you going online and picking things, a chat agent could do that for you, when you texted them.
“One of the founders believed in AI, but the other two founders did not believe that AI could do this. I said we could; I wasn’t just defending myself but was defending the field of AI. We built a great system, but the investors didn’t see enough market, and the company was acquired by Yatra,” Aardra says.
She worked at Yatra for a while, but she had been bitten by the startup bug. She says when you’re given a lot of responsibility you work in a startup, which becomes “addictive”. Yatra was also a startup, but it was a bigger organisation and Aardra felt “there wasn’t much to do in terms of AI”.
Roommates turn friends and co-founders
By this time she had met Ranjana Nair, the Co-founder and CEO of RIoT, through Sulekha, a digital platform for local services. The duo soon became roommates and friends.
“We met because of Sulekha and our friendship grew. Ranjana was running a company with another co-founder and it was called 21st Century OTC, which created interactive surfaces that can communicate with customers. It was built for retail spaces,” Aardra says.
Sanchi Poovaya, the other Co-founder of RIoT, was working with Ranjana then. Aardra enjoyed seeing women entrepreneurs work and hustle; she often spent time with them and even worked on a few projects. One of the projects was to use AI to talk to customers for Wipro.
“I remember being very nervous working on it and showing it to a huge organisation like Wipro. About 30 engineers were waiting to see the demo. I kept wondering if I had created something that they all wanted to see,” Aardra recollects.
It was the first step towards the three of them working together. By December 2015, Aardra left Yatra and began working on different startup ideas. The first idea was to create a wearable for women’s safety, one that would alert authorities.
“We have all been in that position where we are wondering if we will reach home safely. We spoke to a few angels; the feedback was that it was an extremely crowded space and funding would be tough. We tried to argue that penetration wasn’t there or else we would have all bought. Few investors are able to understand the issue of women’s safety from a woman’s perspective,” Aardra says.
A problem that hit home
The trio then realised the need to create something original to get funding. They hit the drawing board, and not long after Aardra’s cousin was diagnosed with epilepsy.
“She was an energetic person who wanted to become a dancer and singer. But this meant that she had to take some drugs that would make her really sleepy. Even then, you don’t really know when an attack can happen; it can be fatal. It was an emotionally tough time. I wondered why wasn’t there anything that could detect her seizure without someone being around all the time?”
There was a wearable, but the cousin didn’t want anything that had to be worn on the body, especially since the problem of battery exploding was then rife.
It set Aardra thinking if there was a way to monitor seizures without putting anything on the body. This was the genesis of RIoT. She began doing research on epilepsy, and the changes it wrought in body and brain functions.
“We realised that epilepsy is a medical condition that can manifest in different ways - there are convulsive and non-convulsive seizures. There are seizures where there is no body movement except for the eyelids. If people were going to buy our solution it should work 100 percent for it is a life-threatening condition. And, so we pulled back,” Aardra says.
Pivoting to monitor respiration
They then thought of sleep and health monitoring, particularly respiration rate monitoring. One of RIoT’s first products, Raybaby, is possibly the first non-contact sleep and breathing monitor for babies. The device won several awards at CES and ABC Kids Expo.
When they first got the idea, they started talking to investors for their first cheque. In 2016, there was a Grace Hopper conference in Bengaluru, and the trio coughed up Rs 25,000 for a ticket so Aardra could attend the event.
“I gravitated to meet the investors, more so the woman investor (Varsha Tagre, MD, Qualcomm Ventures). I went to that event to talk to her and get her feedback, and she didn’t come. I decided to write a cold LinkedIn message to her and to my surprise she responded. She asked us to come to their event at Ritz Carlton. It had all the Asia Heads and I went in petrified. I pitched to Varsha in person and was shivering. I think she saw this nervous kid, and she was very friendly and kind, and I will never forget that,” Aardra says.
She says that an early bit of kindness goes a long way. “We were in our early days and dreaming about a future that could or could not happen.” Aardra was introduced to Pi Ventures’ Manish Singal, and the founders learnt that a prototype was needed.
They built that in 2016 and hit another conference, which was giving free tickets to women entrepreneurs.
“We pitched to a lot of investors and kept getting a no. And at that time we didn’t realise that we were doing one of the hardest things - building a hardware startup. It is deeply discouraged; software seems cooler in terms of returns,” Aardra says.
Fortunately, they had Arpita Ganesh, Founder of Buttercups, as an advisor. She introduced them to the founder of a wearable startup, who asked them to pitch to Benjamin Joffe of Hax Ventures.
“We went and showed a prototype to him and got into a call with a venture partner in China. They loved the idea and signed our first cheque of $100,000. I honestly think we wouldn’t have raised that money in India because hardware is a big no-no,” Aardra says.
RayIoT is supported by SOSV, Anthill Ventures, Madison Ventures, and HCG Hospitals. The startup has a 20-member team. Now, the trio has built a non-contact, wi-fi enabled, affordable respiration rate monitor for coronavirus patients; this is called RayIoT and can run as mini ICU units.
It raised $1.5 million in 2018, and is in the middle of raising the next round.
Aardra still codes and builds the core systems. For her, initiative is the most important thing. “How much are you willing to go that extra mile is what I look at.”
Edited by Teja Lele