Team 67_CareerSpire

Inspiration

The modern job hunt for a fresh graduate feels exactly like a "roguelike" game: the path is unclear, the challenges are randomized, and facing the "Final Boss" (the interviewer) without the right equipment (skills) leads to immediate failure (rejection).

We looked at the problem statement, and realized that existing solutions like LinkedIn or MyCareersFuture are just "lists." They lack engagement and progression. Also, the career data all over the internet makes it difficult for the individual to piece out and plan an actionable roadmap to reach employment.

Inspired by the game Slay the Spire, we wanted to transform the terrifying ambiguity of finding a job into a structured, gamified adventure. We wanted to build a platform where upskilling isn't a chore, but a questline, and where the "loot" at the end of the dungeon is a real-world career opportunity.

What it does

It's a gamified career roadmap platform that turns the job search into a strategic adventure. It asks for the user's resume and offers two distinct gameplay modes:

The Specialist Path (for Decisive Users): For users who know their target (e.g., "Full Stack Developer"), the system generates a procedural map.

The Nodes: The user traverses a map of "nodes." Each node represents a specific upskilling task or milestone (e.g., "Learn React Hooks" or "Build a Portfolio Project") derived from the Jobs-To-Skills algorithm.

The Boss Fight: Once the nodes are cleared, the user faces the "Boss": an AI-powered Mock Interview tailored specifically to the job description.

The Loot: Defeating the boss unlocks a "Chest" containing a direct application link to real jobs related to the user's specialization. (sourced from competitive datasets such as Glassdoor and MyCareersFuture).

The Explorer Path (Indecisive Users): For fresh grads who are lost, we built a discovery engine.

Personality & Asset Analysis: The user takes a personality test, which we cross-reference with their resume data.

Branching Fates: The system recommends 5 distinct "Class Specializations" (Career Pathways) suited to their profile.

The Journey: Once a path is chosen, they enter the same node-based progression system to build the skills they were previously missing.

How we built it

We engineered a full-stack "Game-ified Career Engine" that transforms static job data into an interactive Directed Acyclic Graph (DAG). Using an Express.js and Supabase backend, we built a pipeline that scores job descriptions with RPG attributes (STR/INT/DEX) and integrated DeepSeek’s API to perform semantic gap analysis on user resumes, generating procedural "Skill Trees" and dynamic "Boss Fight" interview personas. The frontend, built with React, Zustand, and Framer Motion, visualizes this journey as a navigable map where users must pass role-specific technical quizzes to unlock subsequent nodes, effectively converting fragmented career information into a verifiable, gamified roadmap.

Challenges we ran into

The hardest problem was the "Cold Start" Dilemma (The Explorer Path). We realized early on that tackling fresh graduates who don't know what they want is significantly harder than helping those who do. Initially, our recommendation engine was too generic. We struggled to find the right "indicator" to gauge a user's capability without a job history.

The Fix: We pivoted from just analyzing the resume (which is often empty for grads) to implementing a Personality & Interest Test. Calibrating this test to meaningfully map to the 5 suggested pathways was a massive design challenge. We had to ensure the "suggested specializations" felt personal and accurate, not just random.

Accomplishments that we're proud of

The "Boss Fight" Mechanic: We successfully turned the scary experience of an interview into a fun, repeatable gameplay loop. It provides instant feedback, which is exactly what the "Personalized Upskilling Roadmap" bonus feature asked for.

Solving the "Analysis Paralysis": By forcing indecisive users to choose 1 of 5 suggested paths, we successfully gamified decision-making, helping them break out of the cycle of uncertainty.

What we learned

Gamification is complex: Balancing "fun" with "utility" is hard. If the game is too long, users will lose interest. If it's too short, they won't learn much.

Interpreting Unstructured Data: We gained significant experience parsing the "Job Description" columns from the datasets to extract actionable keywords for our node generation.