Team38_HireMeLah!

HireMeLah! is a unified career platform for fresh graduates finding jobs. It quantifies how well a role fits you, tracks your applications automatically, and generates personalized learning roadmaps.

Comment

About the Project

HireMeLah! is a web-based application designed to help students and early-career professionals better understand how well their skills align with specific job opportunities. The project was inspired by a common frustration in the job search process: resumes are unstructured, job descriptions are often vague, and candidates are left guessing whether they are a good fit. HireMeLah! aims to make this matching process more transparent, data-driven, and interpretable.

Motivation and Inspiration

Many job-matching platforms rely heavily on keyword matching or opaque ranking algorithms, which can be misleading and difficult for users to trust. I wanted to explore whether it was possible to:

  • Translate resumes and job descriptions into structured signals
  • Quantify fit in a way that is interpretable rather than purely predictive
  • Present insights that help users understand both their strengths and areas for improvement

This led to the idea of combining AI-powered parsing with simple, interpretable indices, rather than relying on black-box scoring methods.

System Architecture and Technology Stack

HireMeLah! is built on a modern, scalable web architecture that supports rapid iteration and future extensibility.

  • Frontend:
    The user interface is built with Next.js 14 and React 18, enabling server-side rendering and fast client-side interactions. Styling is handled using Tailwind CSS, which allows rapid UI iteration while maintaining consistency.

  • Authentication and Integrations:
    User authentication is implemented using NextAuth.js with Google OAuth, providing a seamless login experience. Gmail integration allows resumes to be securely accessed directly from the user’s inbox.

  • AI and Resume Parsing:
    Resume parsing and information extraction are powered by OpenAI’s GPT-4o-mini. The model is used to convert unstructured resume text into structured representations of skills, roles, and experience.

  • Type Safety:
    The entire application is written in TypeScript, ensuring end-to-end type safety across frontend, backend, and API boundaries.

Methodology and Scoring Framework

At the core of HireMeLah! is a transparent scoring system designed to balance simplicity and usefulness. Instead of producing a single opaque score, the system is built around three interpretable indices.

Skill Fit Index (SFI)

The Skill Fit Index measures how well a user’s skills meet the requirements of a job listing.

SFI = (1 / N) × Σ min(1, Y / X) × 100

Where:

  • N is the number of required skills
  • X is the required proficiency level
  • Y is the user’s estimated proficiency level

This formulation caps overqualification and focuses on requirement coverage rather than excess skill depth.

Relevance Index (RI)

The Relevance Index measures how closely a job aligns with the user’s intended career direction.

RI = 100 × (0.50 × Category + 0.25 × Role + 0.25 × Skills)

This weighting prioritizes overall career category alignment while still accounting for role and skill relevance.

Career Fit Index (CFI)

The Career Fit Index combines both dimensions into a single interpretable score. CFI = (SFI × w) + (RI × w)

By default, both indices are weighted equally (w = 0.5), though the framework is flexible and can be adjusted in future iterations.

What I Learned

Through this project, I learned:

  • How to design and implement a full-stack web application using modern frameworks
  • How to integrate AI models responsibly as decision-support tools rather than decision-makers
  • The importance of interpretability when applying statistical or AI-driven methods to user-facing products
  • How to translate qualitative concepts, such as “career fit,” into structured and testable metrics

Challenges Faced

One of the biggest challenges was converting unstructured resume data into consistent, comparable representations without losing nuance. Another challenge was avoiding overfitting or false precision when assigning numerical scores to inherently qualitative information.

Balancing technical sophistication with clarity was also non-trivial. The system needed to be robust enough to handle diverse resumes while remaining simple enough for users to trust and understand the results.

Conclusion

HireMeLah! demonstrates how AI, modern web technologies, and transparent statistical thinking can be combined to improve the job-matching experience. Rather than replacing human judgment, the platform aims to support it by helping users better understand their strengths, gaps, and alignment with potential career opportunities.

Built With

Share this project:

Updates