Tanmay Bakshi is poised and confident standing center stage in a Walmart Technology conference room in Bentonville, Arkansas, like a veteran speaker about to recite a well-rehearsed keynote. He’s introduced to the audience of seasoned developers who’ve gathered to hear him speak, but he needs no introduction; his reputation precedes him. Over the next half hour, his presentation will cover a breadth of topics, ranging from the Ask Tanmay app — the world’s first Natural Language Question Answering system powered by IBM’s Watson — to the artificial neural network he built to mimic human gameplay in a side-scrolling iOS platformer called Nimble Ninja.
Bakshi is 13 years old, a fact that might elicit some questions as to the depths of his engineering prowess. But make no mistake: He’s been at this for quite a while. He’s in town today as a guest of IBM, visiting from his native Canada to speak to students at Bentonville High School. Word of his visit quickly reached Adam Holland, a senior director at Walmart Technology, who extended an invitation to tour the forensic services lab, which eventually led to Bakshi delivering a last-minute talk in the conference room.
Bakshi says he first forayed into the world of coding nearly a decade ago.
“When I was 5 years old, I didn’t really have much to do. Since my dad worked as a computer programmer, I would watch him programming almost every day. It really fascinated me how a computer could display my name on the screen or add two numbers, or really do anything. At 5 years old, that was like magic to me,” Bakshi recounted. “I wanted to find out a bit more about how that actually worked.”
He began with the fundamentals, studying languages like FoxPro and Visual Basic and using the internet as a learning resource to master other core concepts. By age 9, Bakshi was already deep into iOS development. That year, his first app, tTables, was accepted into the iOS app store. Since he was a fourth-grader and not yet of legal age, the app was published under his father’s name.
Bakshi explains that an early driving force in his relentless pursuit of coding was the fact that he simply wasn’t familiar with the concept of a job yet. He didn’t know programming was something people were paid to do, so he never considered it work. It was just another toy. That attitude continues to serve him well. Today, Bakshi approaches cutting-edge technologies like machine learning and artificial intelligence with the same youthful exuberance his peers might bring to a new PlayStation game.
Aside from likely being the youngest developer to ever work with Watson, Bakshi is also one of the platform’s most fervent evangelists. Through his YouTube channel —which boasts more than 159,000 views and 4,500 subscribers — Bakshi offers tutorials on how to use Watson for everything from identifying dog breeds to controlling an AR drone. He frequently fields questions from adult developers and even companies seeking to implement Watson services in their own offerings.
But Bakshi’s tutorials go well beyond Watson explainers — ranging from the basics of coding in Swift to the fundamentals of blood typing. He says his goal is to help at least 100,000 aspiring young technologists gain a foothold in STEM disciplines and estimates he’s already reached 2,400 people. Through his YouTube channel, keynote speeches and soon-to-be published book, he hopes to reach many more.
In the meantime, Bakshi looks ahead to what’s sure to be a fruitful collaboration with Walmart Technology. Specifics are still being hashed out, but he’s set to work with the company’s data analytics team and hopes to be back in Bentonville in the near future to present a new social media analytics tool he’s been working on.
Walmart has also assigned Holland as a mentor: Holland says he was eager to take on the role with one caveat: that Bakshi will probably learn less from Holland than the mentor learns from him. But Bakshi laughs off that notion.
“I actually have a lot to learn from Walmart,” he said. “Walmart is working with technologies like neural networks and deep learning, and I really believe these sorts of cognitive computer technologies are vitally important for both the present and the future.”