Finding the Hidden Revolutionaries

Revolutionaries in Squid Game For this project, I used Google Gemini to help identify a hidden group of “revolutionary” players in a fictional Squid Game-style competition. These players had to be extremely intelligent and have a background in combat. That meant I needed to sift through a massive dataset full of participant bios, IQ scores, and background stories — many of which were vague or full of creative job descriptions. Here’s where Gemini came in. I leaned heavily on its natural language understanding and data analysis features. The trickiest part was the “combat background” bit — some bios mentioned things like “military drone technician,” while others said things like “experience in the ring as a fighter” or “security detail abroad.” I used Gemini to help read and interpret these descriptions like a human would, picking up on keywords, phrases, and context clues that pointed toward any combat or tactical experience. For example, one prompt I used was: “Review this background text and tell me if this person likely has real-world combat experience. Be strict and look for key phrases like military, police, special forces, etc.” Gemini was surprisingly good at this. It could distinguish between someone who just worked for a defense contractor and someone who had boots-on-the-ground military experience. Once I had a filtered list, I used Gemini again to help prioritize candidates based on IQ and other traits. I chose Google Gemini because I needed something that could read between the lines. A regular keyword search wouldn’t cut it — the descriptions were too varied, and the context mattered a lot. I needed a model that could interpret language in a human way but faster and more accurately. The idea came from playing around with the concept of identifying “hidden threats” or rebels in a system — like finding needles in a haystack using language clues. I thought, what if I could build a tool that flags people based on subtle traits and text patterns? That’s how the idea evolved into a Squid Game-style selection problem. My approach started with trying simple filters in code (like IQ >= 140), but when it came to the backgrounds, I hit a wall. That’s when I brought Gemini into the loop. I experimented with different prompts, tested how it responded to creative versus straightforward bios, and kept tweaking things until I got confident, consistent results. What makes this project stand out is how Gemini wasn’t just generating text — it was reading and analyzing it in a meaningful way. I used it to interpret vague or euphemistic language and uncover hidden intent or experience. That’s not something you can do easily with traditional filters or regex matching. Also, I think the concept itself is pretty fun and creative — using an AI model to help run a dystopian game show, but with logic and analysis driving the decisions. It’s like mixing psychology, storytelling, and data science, all powered by Gemini. This project pushed the boundaries by using Gemini almost like a human teammate — someone who could read between the lines, give feedback, and help make judgment calls. It felt like I was working with a partner, not just a tool.

Built With

• python