Rahul Joshi is a man who knows his stuff. His current role as Software Development Manager at Amazon proves this. Listening to him talk about software and search systems makes you realize the difference between a subject expert and one who isn't. It may not seem so at first glance, but Joshi's expertise comes from years of experience in the field. And that journey started as early as his high school days.
Early Engineering Days
We all have our extracurricular activities after school lets out. For Rahul Joshi, it was science and mechanical engineering. Carpentry, smithery, electronics, and the science behind them fascinated Joshi.
"I've always been interested in how the world around me works," Joshi says, recalling his younger years with a laugh, "Influencing this world by building all these things just drew me in. I remember attending these science quiz competitions every so often."
This early interest set the foundations for professional work.
Pivoting to Digitization
"Mechanical engineering requires machines available only in the labs," Joshi explains his dive into learning programming languages, "At first, it was convenience."
He talks about teaching himself several popular programming languages alongside working databases.
"All machine designs were done using 3D modeling software," he says, talking about his time working with automobiles. Joshi worked with software so much that he began taking a deeper interest in digitized simulation and the coding behind it.
"The creativity and innovation was suddenly more appealing to me," he says, grinning.
His interest led him on a whirlwind career journey in which he worked on all kinds of software. He created and developed code, starting his own company and facing its ups and downs, coming out of it stronger than ever.
Eventually, he arrived at Amazon, one of the biggest names in online shopping today.
Why Amazon?
Joshi's experience with start-ups is what led him to Amazon's metaphorical door.
Start-ups have a specific energy in their culture and innovation. And Joshi found that Amazon's own 'Day 1 Culture' matched that energy.
"Amazon's is the definition of large scale as both a company and in its customer base," Joshi says, "I saw it as an exciting opportunity to challenge myself with large scale operations while also expanding my problem-solving and creative skills."
After joining, Joshi's hopes were met head-on. He worked with his colleagues to improve software, quickly execute solutions, and experiment with iterative enhancements.
"The way the work culture supports individual growth within a strong support system resonated with me," Joshi says, thinking back, "It made me want to pay it forward."
A Culture of Development
Perhaps what differentiates Amazon from any other Fortune 500 company is their culture.
One approach that might be familiar to people in management is the 'two pizza team,' coined by Jeff Bezos. It emphasizes "the importance of small, agile, and autonomous teams" that can "promote efficiency, effective communication, and a sense of ownership within each team" [1].
"Team independence allows us to execute plans and support customers easier," Joshi explains, "I have personally found it increases the amount of creativity that arises in both individuals and the team as a whole."
However, independence needs to be tempered by connectivity to the company's goals. Something that Joshi says he experienced in abundance at Amazon.
"We would do deep dives into data and metrics during business meetings," he says, his seriousness belied by an easy-going grin, "We took every person—a colleague or customer—seriously, and made data-backed decisions."
He says their bottom-up approach allowed them to fix problems efficiently and develop a system where goal-setting and proposals were encouraged.
Search System Satisfaction
"The number 1 rule in developing search systems is customer satisfaction," Joshi explains, "It creates more business opportunities and drives what I call customer delight opportunities."
For a company as large as Amazon, dealing with issues in their search engines is time-sensitive. Making their search systems faster and more efficient is an overarching goal that can never be forgotten. Getting the customer that one record they want out of a billion is so important for customer satisfaction, retention, and further engagement.
"We look at our current metrics and keep digging to see where we are lacking," Joshi says, talking about how his teams enhance searches, "We look at how new technology fills these gaps, and how quickly it manages to do so. Speed is everything."
Before working with search engines as a whole, Joshi worked on Amazon's recommendation algorithms. One example he expounded on was movie recommendations. Data metrics such as how many times a recommended movie was chosen or issues with recommendations not matching the user-selected criteria were his bread and butter.
Joshi talks about the concept of 'customer affinity,' where customers' personal preferences are derived from their behavior while using search engines, and how he has used techniques specific to search systems to enhance customers' experiences.
"You can say it's another example of our Day 1 Culture," Joshi explains, "Be it a problem with search, entity resolution, or quality, how it affects the customer is the key detail. We ask ourselves why and how a customer feels, and work based on that."
Moving Up the Ranks
Joshi first joined Amazon as a Software Development Engineer, becoming a Senior Engineer several years later. But just two years after that promotion, he was elevated to the position of Software Development Manager.
When asked about the differences between the roles, his answer was his scope of influence.
"Previously, I influenced my team and vice versa," he explains in-depth, "You get guidance from your seniors and there's not much ambiguity in the strategies and solutions chosen."
However, as a manager, he found himself in charge of building high-performance teams. He was the one guiding his juniors, facilitating communication and collaboration while also managing stakeholders and inter-department communications.
"I found my focus shifting onto people," he says, "Managing and coaching those I was in charge of on one hand. And maintaining that extreme focus on customers on the other."
Of course, these tasks were not without challenges. But he worked with them anyway. "It may seem simple, but it's about delegation. Or more specifically, delegating to the right people."
Joshi talks about breaking up big tasks into smaller ones and understanding their complexity. "Who can handle what?" he posits, "How can I facilitate career growth while maintaining a project's goals? I had to look at the macro scale while also following up with the people right in front of me."
The Future of Search
While there aren't enough pages to encapsulate the entirety of Joshi's knowledge of search systems, his opinion on its future is a good place to start.
"There is a drastic change happening in the world of search systems," Joshi explains, "It's mostly due to LLMs and the RAG approach."
Large Language Models and Retrieval Augmented Generation seem to be pushing towards 'Semantic Search Systems,' leaving older models behind in the dust.
"Technology is not the only factor in this change," Joshi says with passion, "We are now seeing systems that work based on the customer's intent, making them more intuitive and conversational."
He explains how earlier systems used keywords, which can be clunky and error-prone for those not knowledgeable about it. Now, the software interprets the words and sentences inputted to generate better results.
"Another aspect I am seeing more often is how search results bring user-specific choices," Joshi explains. He uses the example of a tennis racket: "In the past, a search for good rackets would give you a number of sources for you to do your own research. Now, the results showcase recommendations and specific choices."
Joshi says the future of search lies in action. This is not just improving the speed of a system's action of interpreting a query but also providing actionable items to users.
"It's about getting things done," he says.
Rahul Joshi's passion and expertise in the field of search systems started from his excitement to influence the world around him. Now, in his prominent position at Amazon, he has the ability to effect change for millions of users. It may seem like his path has reached the peak of the mountain, but he is still rearing to go further beyond.
