Inspiration
The inspiration for our project came from observing how confusing and fragmented the job-search process is for students and fresh graduates. Career information is spread across multiple platforms, job descriptions are often unclear, and it’s difficult for individuals to understand how their current skills align with industry expectations or what they should focus on learning next. As a team, we felt that many people have strong potential but lack clear direction. Since this was our first hackathon, we were especially motivated to work on a problem that felt real, relatable, and impactful.
What it does
Our project is an intelligent job-matching and career-guidance system centered around a user’s CV. The system extracts skills from a CV, identifies the job category that best fits the user based on those skills, and matches them with relevant job opportunities. It then performs a skill gap analysis to show what skills the user already has and what they need to improve or learn. Based on this, the system generates a personalised, time-based learning roadmap and continuously updates recommendations as the user tracks their progress. Instead of being a one-time job search tool, the platform acts as a long-term career companion that helps users grow into the roles they aspire to.
How we built it
We built the system as a step-by-step pipeline that turns a CV into actionable career insights. Resumes in PDF or DOCX format are parsed using pdfplumber and python-docx, and key sections like skills, experience, and projects are identified for accurate extraction. Extracted skills are matched against a master skill catalogue and classified into technical and soft skills. A machine learning model then predicts the most suitable job category based on skills rather than job titles. Relevant jobs are fetched using an API, and an LLM generates a personalised learning roadmap by comparing the user’s skills with industry requirements. As this was our first hackathon, building and integrating an end-to-end system was a major learning experience for the team.
Challenges we ran into
One major challenge was dealing with unstructured and inconsistent CV data, as resumes vary widely in format and wording. Skill extraction required careful normalisation to handle different representations of the same skill. Ensuring that job categorisation remained skill-based and unbiased by job titles was another challenge. As first-time hackathon participants, managing time, coordinating tasks, and debugging under pressure were also new experiences that pushed us to adapt quickly and work more efficiently as a team.
Accomplishments that we're proud of
We are proud to have built a complete CV-to-career guidance pipeline within the hackathon timeframe. Despite this being our first hackathon, we successfully integrated resume parsing, skill extraction, machine learning-based job categorisation, job matching, and personalised roadmap generation into a single working system. We are especially proud of creating a solution that goes beyond job recommendations by guiding users on how to become qualified for their desired roles.
What we learned
This hackathon taught us that real-world data is messy and requires thoughtful preprocessing before meaningful analysis is possible. We learned how powerful APIs and LLMs can be when used thoughtfully, especially in transforming raw data into clear, actionable guidance. As a first-time team, we also learned the importance of communication, task distribution, and iterative problem-solving in a fast-paced environment. Most importantly, we gained confidence in our ability to learn new tools quickly and combine machine learning, APIs, and LLMs into a practical, user-focused solution.
What's next for Team77_J.F.Y
Moving forward, we plan to enhance the platform with real-time labour market APIs and more adaptive, LLM-driven personalisation. Beyond technical improvements, this first hackathon taught us strong lessons in time management, teamwork, and working under pressure. As a team, we are now better prepared to plan efficiently, take on bigger challenges, and perform even better in future hackathons while developing this project into a scalable, real-world solution.
Log in or sign up for Devpost to join the conversation.