Five engineering students from BITS, one of India’s most prestigious institutes, decide to make the most of their Jivox internship by diving headfirst into the company’s hackathon. What could go wrong? Well, as Team LLMinators would soon discover, sometimes your AI develops an inexplicable obsession with seasoning.
As current BITS students, Amarjeet, Krishna, Tarun, Sankeerth, and Shubham naturally gravitated toward each other when people started to form teams for the Jivox Labs Hackathon. “We’re all from BITS, so we just clicked,” they explained, exhibiting that quintessential college spirit.
The team chose to tackle the recommendation system challenge, aiming to help brands automatically identify the best-performing products for specific audiences.
Building an AI that could intelligently match products with target demographics seemed simple enough. But they had no idea that the two weeks ahead of them would put their endurance, coding prowess, and their tolerance for some pretty salty humor, to the test.
The Great Salt Incident Of 2025
Everything started promisingly. The team dove deep into preprocessing, meticulously defining table structures and column meanings. But then came the moment that would cause the team to rip their hair out.
“It was recommending salt for everything,” they recalled with a mix of horror and laughter. Athletes looking for shoes? Salt. Tech enthusiasts browsing gadgets? Salt. Beauty influencers seeking cosmetics? Yup, you guessed it. Salt.
Imagine telling prospective customers: “Well, our sophisticated AI analysis indicates that we should promote salt. Yes, to absolutely everyone.” It was like having a recommendation engine with the dietary preferences of a medieval physician. After a point, the AI engine seemed to invoke that Oprah meme (you know which one): “You get salt…you get salt…everyone gets salt!”
The Iteration Marathon
To their credit, the team didn’t panic. Instead, they rolled up their sleeves and began what can only be described as an iteration marathon. They first tried to design and train their AI from the ground up, believing they could produce something groundbreaking. Halfway through, reality set in, and they saw their strategy was fundamentally flawed.
“We realized we were overcomplicating things,” they admitted. The team regrouped, pivoted to leveraging existing AI models more effectively, and began fine-tuning their confidence scoring system. The team had one burning goal: if an audience had nothing to do with culinary seasoning, stop recommending salt!
The Final Sprint
As presentation day came, everyone was sprinting towards the finish line. Even with the extra hours spent in the office, the team was fueled by the type of optimism that comes from knowing you’re onto something good. Even a few hours before their presentation, they were still tweaking and improving their solution!
“We were literally experimenting right until the end,” they laughed. “And it was only much later when we realized our frontend wasn’t even integrated!”
While this may appear to be a nightmare for those with a high Conscientiousness score in their Big 5, sometimes the best solutions emerge from controlled chaos. LLMinators managed to fix that pesky salt-loving bug and created a simple click-based interface that eliminated manual audience matching.
Their solution will help marketers reduce campaign creation time. Now that’s a pretty savory outcome after all those salt recommendations!
Different Folks, Different Strokes?
The LLMinators’ journey illustrates how innovation thrives on different approaches. While some teams might prefer methodical, structured strategies, this group proved that creative chaos, when channeled correctly, can yield amazing results. Their experience reflects the broader spirit of innovation at Jivox Labs, that drives products like IQ DaVinci, where diverse thinking styles and constant experimentation make for solutions that look to solve real-world problems.
As second-place winners, Team LLMinators proved that sometimes the best path to success involves things getting a little salty along the way. Their recommendation system now powers through data with intelligence and precision, helping marketers optimize campaigns without any unwanted seasoning suggestions.
The future looks sweet for these five engineers, who turned their internship into an unforgettable learning experience. After all, any team that can debug salt-obsessed AI and still come out smiling clearly has what it takes to tackle whatever challenges come next.