[Techie Tuesday] Meet Archie Agrawal, who went from Indore to Seattle to build systems for BofA, Microsoft, and Amazon
For Archie Agrawal, getting into engineering and tech was anything but the norm.
Hailing from Indore, Archie’s father is a chartered accountant and mother runs a beauty salon. Technology didn’t really have a role to play.
But destiny had other plans.
Archie was five when she learnt Logo on a computer at school. And, there was no looking back.
Today, Archie is the product lead at Amazon in Seattle. Over the years, she has conducted research in areas such as requirement engineering, logical formalism, and semantic web concepts. She also developed a tool to make domain knowledge visible and accessible for reuse and reconfiguration, and tracing.
At Amazon, Archie is working to build human-in-the-loop machine learning for managing user-generated content.
“When I first used Logo to draw basic shapes and fill colour when I was five, it wasn’t a big thing. However, it was a powerful moment for me…I felt I could create things using just a keyboard and a computer. It was extremely empowering,” she says.
A few years later, she saw her father’s office had computers, which he would use for accounting. “He would let me play video games and explore the system,” Archie recalls.
Engineering and research
By the time she was in Class 10, Archie had a strong calling towards a field in engineering and computer science. After Class 12, she joined SVITS, an engineering college in Indore, to pursue a bachelor’s course in computer engineering.
“I learnt different languages in those four years. I was sure I wanted to do something in computer science, and wanted to get into a top university. Since I didn’t get into one in 2013, I decided to do some research work before reapplying,” she says.
In 2013, Archie joined as a software engineer at Tata Research and Development Centre in Pune. Here, she researched in areas of semantic web concepts, requirement engineering, and logical formalism, which is derived from knowledge extraction, representation, domain ontologies, and reuse and reassigning using different NLP techniques.
Archie built a tool to make domain knowledge visible and accessible to make it easier for configuration and reuse. It also helped in analysing regulations from an implementation perspective and in tracing regulations to the requirements.
In 2014, she moved to Pittsburgh to purse a master’s course in computer science at Carnegie Mellon University.
“It was a life-altering experience. My university gave me the opportunity to dream big while staying rooted in reality, and build products. I didn’t realise how intense a master’s programme at CMU would be. I would pull all-nighters at the Hunt Library. Also, you are living on a shoestring budget as a student in a different country,” Archie recalls.
Understanding ML and Carnegie Mellon
During this time, Archie felt the need to learn machine learning (ML) from the best professors in the industry. She ended up taking ML classes that allowed her to understand the nitty-gritties of machine learning models and how they work behind the scenes.
“I had an opportunity to partner with NASA to build an evaluation platform for efficiency, cost, and ability to meet security and compliance,” she says.
She was also a part of the Yahoo! InMind project where she built a rapport recogniser, a classifier with non-verbal and verbal features as inputs and output rapport levels.
During her three-month summer break in 2015, Archie interned at Schlumberger, one of the world's leading oilfield services providers, in San Francisco.
At the oil and gas company, she built a voice-enabled search for oil sites so that people working on heavy machinery didn’t have to rely on hand-held devices for information.
The world of finance
After graduating in 2015, Archie joined Bank of America.
“One of the company’s strategic initiatives was to stop and prevent fraud in digital banking. It felt like the perfect opportunity for me to apply my machine learning skills. I was part of the pioneering team that used cutting-edge technology to detect fraud and keep the customer’s money safe,” she explains.
She adds that this was one of her the biggest learning experiences, as it was at “a really large scale”. The work was in the financial space, with a huge opportunity to leverage data for machine learning she learnt to build for scale.
“Building models in the financial space comes with its own challenges. In addition to building real time, you also have to deal with regulatory and compliance requirements. One thing that stood out was how having a deep understanding of customers - both good and bad - could help us tap into signals and data creatively,” Archie says.
Bank of America has nearly over 40 million active digital banking users across web and mobile. This translates to one in seven American consumers and more than 1.6 million logins just on mobile in a quarter. Amidst this scale, fraud detection would happen real time.
The scale of Microsoft
After two years, Archie felt the need to expand her horizons beyond banking, and in 2018 joined Microsoft.
“At a company level, Microsoft aims to make every human more productive. And with digital transformation, they believe that productivity and commerce has to happen through interconnected apps and services that we have access to through an ever increasing number of devices,” she says.
This means developers have to create applications that should be able to connect people and services for collaboration while keeping their focus on the core value proposition. The question was how to do this while keeping users and their data secure.
“I used to work on the Microsoft identity platform, which was a service on Azure for everything identity management. It handled the complexity of authenticating and authorising users. I created a strategy for a developer-friendly experience on a thriving application ecosystem built on the Microsoft identity platform.”
Archie explains this was also a large-scale platform, with more than a million users using the Microsoft identity services. “I used to wake up feeling that what I do really mattered because reliability would be hit if one of our services didn’t function properly.”
The Amazon journey
The desire to keep learning and growing continued to drive Archie, and two years later she felt the need to understand the consumer space to get a “rounded experience”.
So, she signed on the dotted line when she was offered a role at Amazon, which she feels is the “most customer-centric company”. Since 2020, Archie has been working on content risk management to minimise customer and business risk associated with harmful, plagiarised, disappointing, or non-compliant content.
“I'm responsible for ensuring Amazon customers have a safe and trustworthy experience, in line with their expectations on the Amazon store across the world,” she says.
Safety is the most important aspect at Amazon, and Archie works on defining content policies and understanding the downside impact when content policies are circumvented. On the tech side, she has worked on policy enforcement which involved understanding if the content should go through the recommendation system or be suppressed.
“At Amazon’s scale we can’t review every single submission manually. Most of the work involves automation, and combining ML and heuristics. It was an opportunity to come a full circle and use ML for social and business impact. We use NLP for new information that is constantly fed, to get better and proactively detect suspicious products,” Archie says.
Giving back and lessons learnt
Archie also actively invests in startups through SeaChange Fund. She says helping other founders do well in their business helps multiply their impact. “I also get to learn about new industries in a short time.”
Working in top companies has reiterated the importance of one crucial lesson; the importance of building products with the customer in mind.
Today, while hiring techies, Archie looks for “curiosity” as she feels a curious mind can learn new things even if they don’t have the skill to do something. She also looks for “grit and perseverance”.
“My favourite question is if you were to do something differently at all positions at different jobs, what would you do? That answer gives me a window into how they think and the mental model they apply. This helps understand what skills are inherent and what they learnt on the way,” she says.
Offering advice to young engineers and techies, Archie says: “Continue to believe in yourself. Go ahead and do it, and don’t let anyone tell you otherwise. Stay curious and keep advocating for the customer.”
Edited by Teja Lele Desai