At the India edition of Amazon Career Day 2021, Anil Warrier, Head, Talent Acquisition, Amazon Web Services moderated a panel discussion on ‘Tech for Good’ which talked about the pivotal role of technology across Amazon.
From the very beginning, Amazon has been an early proponent of technologies such as Machine Learning (ML), Artificial Intelligence (AI), Alexa, Cloud , many more.
“Amazon’s journey for close to two decades in India has been made possible by the huge contributions from employees we’ve hired.”
Speaking about the role of employees in driving these technologies, Anil said, “Amazon’s journey for close to two decades in India has been made possible by the huge contributions from employees we’ve hired, who are led by their desire to deconstruct, create, and innovate on behalf of our customers.”
Machine Learning to deliver superior experiences
“Machine Learning is integrated so well into the system that it works seamlessly in the background, such that customers are not even aware of it, but it is an essential part of the entire ecosystem,” said Nikhil Rasiwasia, Senior Manager, Applied Science at Amazon India.
At Amazon, Machine Learning can help deliver highly personalized experiences, resulting in improvements in customer engagement. For instance, digital shopping is still gaining traction in India with fluctuating internet speeds being a challenge, the team has developed models to predict customers who are on a slow or spotty network based on criteria like device characteristics, cell tower information, and the latency of the last request. These customers get an adaptive experience and are served streamlined pages with a lower number of widgets that are easier to navigate.
“We are constantly looking for newer ways will continue to double down our efforts in developing technologies to continuously innovate.”
In search too, Machine learning helps understand query semantics, identify phrases and brand mentions in the query, correct spell errors, support transliterated and colloquial usage, and account of ‘popularity bias’. Amazon’s search ranking algorithms display relevant products while concurrently optimizing multiple other objectives: delivery speed, product quality, price competitiveness, etc.
During the pandemic too, ML teams at Amazon have been heads-down focused on innovating at scale with tech-based solutions, to respond to the new normal. “One very interesting example is the COVID Risk Score Prediction (CRISP) mobile app which we developed to detect and tackle the spread of COVID-19, to provide a safe work environment for our employees in fulfilment centres,” said Nikhil. “The app uses bluetooth signals from mobiles to track social contacts between users generating an infection risk score for every employee. So, if we have associates with very high infection risk scores, we can prioritize them for testing.”
AWS has made it extremely convenient to enable processing at scale, but at reasonable costs such that it automatically starts the machine whenever the task needs to be done and shuts it down once completed to ensure optimization of resources.
“While we have made great strides in this direction, we are constantly looking for newer ways will continue to double down our efforts in developing technologies to continuously innovate and build the best shopping experience for customers by working backwards from their needs,” said Nikhil.
Leveraging the power of voice
“Voice is a big part of the future. We believe that voice will fundamentally improve the way Indians will interact with technology. It can make the complex simple. It’s the most natural and convenient user interface,” said Dipa Balakrishnan, Senior Manager, Alexa Experience at Amazon India.
Talking about the growing popularity of Alexa in India and challenges involved in creating an optimal experience for diverse Indian customers who come in equal numbers from metros like Delhi, Mumbai, Bangalore as well as smaller cities like Spiti, Bundi, and Gadchiroli, Dipa said, “India is unlike any market; language changes every 100 kms. And we have many accents. So, when we first launched four years ago, it was important for Alexa to understand the phonetic nuances and context. We had to account for these nuances, unlearn things we knew and innovate on several aspects.”
“Voice is a big part of the future. It can make the complex simple.”
With the combination of machine learning algorithms and computing power, the team worked to ensure far-field speech recognition works with high accuracy in several noisy environments. Much of this needed to be tuned and set up right for the Indian context.
Dipa also spoke about the Smart home feature on Alexa, the growing popularity of which is creating momentum and opportunity for developers and device makers. “We’re consistently hearing from Indian brands and developers that they want to engage with users via voice and provide more personalized and immersive experiences,” said Dipa. “The most popular skill categories that users ask for are astrology, Bollywood, cricket, devotion and entertainment. In the recent time we have also seen an increase in health-related content such as workout and yoga.”
Alexa in India started with 10,000 skills in October 2017, and now has over 30,000 skills, including over 1,000 Hindi skills- from radio, to quizzes to utilities built by large enterprises, small and medium businesses and independent developers.
“Voice interactions lower the barrier to embrace technology and we will continue to innovate on behalf of our customers to make Alexa more relevant and useful,” said Dipa.
AWS engagement with start-ups and government
AWS’s engagement with India-based startups in the public sector began in 2019 with a clear focus on helping government customers accelerate their missions. Recently, AWS launched the AWS Startup Ramp, an acceleration program that will enable early-stage startups to develop innovative solutions for public sector customers, including but not limited to customers in national and local government, space, defense, and healthcare. “During the development phase of the AWS Startup Ramp program, the team worked with several startups such as agritech startup TraceX and Unmiti, and fintech startup Whrrl, focused on the distinct needs of the government technology sector,” said Kanishka Agiwal, Head, India & South Asia, Service Line, Amazon Web Services (AWS).
“The team worked with several startups that focused on the distinct needs of the government technology sector”
Last year, AWS collaborated with NITI Aayog to establish a Frontier Technologies Cloud Innovation Center (CIC) that addresses their core mission of identifying and deploying leading edge technologies to drive continuous innovation in delivering citizen services and accelerating public sector innovation in sectors such as agriculture, health, skill development and employment. This year, NITI Aayog under the aegis of the CIC, collaborated with the Digi-Yatra Foundation for the Digi Yatra Central Ecosystem (DYCE) challenge for Indian startups. As part of the NASSCOM task force, AWS helped develop a COVID-19 data platform for the government of Telangana in 2020. The solution deploys more than 100 dashboards using anonymized government and public datasets, with hundreds of thousands of COVID-19 related data points.
eGovernments Foundation collaborated with AWS to develop the National COVID-19 e-Pass solution within 72 hours to provide a digital pass management system that supports State Governments of Haryana, Punjab, Delhi, Telangana, Puducherry, Odisha and Karnataka, to manage the lockdown while ensuring smooth movement of the essential service providers across and within states. Earlier this year, the Ministry of Electronics and Information Technology (MeitY) in India established a Quantum Computing Applications Lab in the country, in collaboration with AWS, to accelerate quantum computing-led research and development and enable new scientific discoveries.
“We are also working with the Common Services Centre (CSC) to drive citizen services at the grassroots level that act as access points for delivery of essential public utility services, social welfare schemes, and services pertaining to healthcare, finance, education, and agriculture, besides providing a host of B2C services to citizens in rural and remote areas of the country,” said Kanishka.
AWS is also working with Project StepOne, a non-profit, volunteer-led initiative that was born during the pandemic to augment government resources with technology, people, and processes to effectively fight against COVID-19. StepOne technology framework is entirely hosted on AWS Cloud, as it offers the best-in-class tools and managed services which are essential for an organization like StepOne.
AWS also supports Mediknit, an online academy marketplace for healthcare professionals. In collaboration with GEM Institute of Laparoscopy and Robotic Surgery, Mediknit provides live surgery streams via the GEM Televersity portal for the continuing medical education (CME) needs of surgeons.
The opportunity at Amazon
Machine learning has many immediate applications in areas like image and video analysis, content and product recommendations, speech recognition, and language analysis. The technology is being applied across virtually every industry, including healthcare, retail, media, financial services, and manufacturing to help businesses better serve their customers and gain significant advantages.
“If you are passionate about technology and understand the critical value in being obsessively customer centric, then Amazon is the place for you”
“Machine Learning is one of the most disruptive technologies we will encounter in our generation. Customers today are applying machine learning to the core of their business,” said Nikhil, “So, if you are passionate about technology and understand the critical value in being obsessively customer centric, then Amazon is the place for you.”
Speaking about the opportunity for young talent to contribute to Alexa’s journey in India, Dipa said, “We have a lot of diverse career paths within the Alexa world. Many of our roles index as much on left brain as right brain – unlike popular perception that this may be a world full of research scientists.” In addition to software development roles, research and ML roles, the Alexa team also has roles for user researchers who help understand what the customers want, voice designers who help create these new interaction paradigms, linguists or language engineers who are working in the language understanding space, as well as creative roles in content & marketing
“Alexa’s journey is just starting and there is so much that we still need to understand and create for the customer – we would love for people who have a true builders mind-set, who can think big, are curious to learn, willing to define and re-define a space to come and join us,” said Dipa.
On the distinctive ways of working at AWS, Kanishka said, “At AWS, we don't mind being called ‘peculiar’. We have our own way of doing things. We are obsessed with customers, we see beauty in simplifying the complex, and we are comfortable with being misunderstood.”
“For people who like to invent, there's no better place to build than AWS”
“That might sound unorthodox, but our unusual approach and our culture which is focused on removing obstacles so builders can build, are part of why our employees enjoy working in AWS. So, for people who like to invent, there's no better place to build than AWS,” he added.
To know more open job opportunities at Amazon India visit: https://www.amazon.jobs/en/teams/in.