If ChatGPT and Photomath had a baby, the new app Quizard would be it.
It’s been a full week since UR seniors and Quizard co-founders Michael Giardino and Achraf Golli launched the study app, which harnesses the power of artificial intelligence (AI) to provide students with explanations and solutions to multiple choice and short answer problems. Since their launch day last Saturday, Jan. 28, Quizard’s first TikTok has racked up 1.6 million views and over 170 thousand likes.
“Honestly, I was playing around with it,” Golli says, recounting a normal night for Computer Science (CS) majors that unbeknownst to both of them, would be the start of something extraordinary. “Mike saw the app and he was like, ‘Ok, we can make this big.’”
Giardino recalls how he and Golli encountered video after video on TikTok of students extolling the merits of ChatGPT — the AI bot that has captured the public’s attention since its release in November last year — in helping students with homework. Generative pre-trained transformer models, or GPT for short, use deep learning and language processing to produce responses that mimic text written by humans, drawing from large amassed libraries of data and computing power to return intelligible output.
While modern-day predecessors have been around since 2018, ChatGPT’s flexible and user-friendly interface combined with its marketing for public use set off an AI craze that would reverberate throughout the years to come. And from that hype, plus a little TikTok scrolling session, the idea was born: Giardino and Golli would combine the brains of modern AI and a scanner, wrapped in the convenience of a handy mobile app.
In the technology arms race for novel, cutting-edge GPTs, the duo decided to throw their hats in the ring. “It seemed like a great opportunity to catch that wave,” Giardino says, to which Golli quickly added, “We knew immediately that it was gonna go viral.”
What distinguishes Quizard from other tools is perhaps its ingenious fusion of AI with a scanning feature, whereas other technology requires users to type in questions instead. Self-dubbed “your AI study buddy” on its TikTok page, the app allows users to take a picture of text-only problems, which its algorithms decipher to spit out an answer or explanation — the new go-to for students stumped on a tricky history question and trivia aficionados alike. While Quizard fares well with text, it’s currently incompatible with math problems. That doesn’t stop users from trying, though. “When you say it doesn’t solve math problems, the first thing people will do is try solving math problems,” Golli laughs. Nonetheless, the team does have plans on their radar for improving that feature. Don’t despair yet, engineers.
Quizard’s success cannot be understated. At just a week old, the app has answered over 60 thousand questions. Climbing to number 32 in Education on the App store on Monday, it even ranked in the top 10 Education apps above names like Mathway and Symbolab in Latvia. This success could not have been achieved without the conviction of the full founding team, which includes seniors Michael Kingsley, Mohamed Ali Manai, and Ben Or-Chen — a close friend at Northeastern recruited for his marketing expertise.
Finding the right people was actually easy, according to Giardino. The co-founders assembled the team through connections in UR’s CS department and Giardino’s club RocLab, a student-run organization that built the platform Join AURA connecting undergrads to research opportunities. Currently, the five-person team (excluding Or-Chen) live together, making for a fun breeding ground of spontaneous collaboration.
“It’s almost bad in a sense because me and this guy,” Giardino turns to Golli, “will be walking into each other’s rooms at 4 a.m. and like, ‘I have this idea and we have to do this now!’” Golli’s reply: “We did this multiple — every day!”
The reason behind the team’s nonstop grind: Like the existing plethora of smart AI models, accuracy is key to delivering user satisfaction, which is their top priority. The downside of generative AI is that it can confidently return wrong answers to unknowing users, and the team is only after the best results. Ever since their project began that night in December to their product launch a month later, the work has been nothing short of exhilarating, toiling hours — workdays over the winter break were more relaxed at six hours daily, but through January the project was consuming 12 hours on top of schoolwork on the regular.
Post-launch, the current challenge for Quizard is moving quickly in an evolving market. Sheer manpower on the developing front is the largest issue, they admit, and the team is looking to onboard more developers after finalizing ownership. As for their marketing crew, a looming question stands: how to go viral again? Climbing the charts again is no easy feat, but the Quizard team is also looking for user sustainability. After their first TikTok exploded online, they had no choice but to stay cognizant of competitors, a fervor that lasted them through a series of 5 a.m. all-nighters. Now, the team is continuously analyzing their data to develop insight not only on acquiring users but user retention and integration of user feedback into upcoming features.
Although Quizard started as a side occupation aimed at helping students like themselves, Giardino and Golli hope to flesh out their passion project into a long-term vision. The future of Quizard is in flux, but it’s still too early to tell, the co-founders conclude. After graduation, Giardino will be working at Slack and Golli is joining Intuit, both as software engineers. As for the other team members, Kingsley will be attending graduate school at Cornell, Manai will be starting at Amazon, and Or-Chen’s post-graduate plans are in marketing.
It’s not an everyday occurrence to build a product from start to finish, but now that Quizard has gained some traction, the pair is realizing how much work is ahead of them. On that list includes plans for migrating the app to the Android interface, as well as making the app more user-oriented through various features coming soon. As Golli says, “We proved the concept — now we actually have to deliver.”
