Hired

Your career, guided by intelligence — not luck

Comment

Inspiration

Every year, thousands of graduates leave university with strong skills but no clear path forward. I watched friends spend months applying to jobs blindly — tailoring CVs manually, getting no feedback, and never knowing why they were rejected. The job market felt like a black box. At the same time, I was deep into courses on Deep Learning, NLP, and Big Data — and I kept asking myself: why isn't any of this intelligence being put to work for the people who need it most? That question became Hired. I wanted to build something that wouldn't just list job offers, but would understand the person behind the CV — their skills, gaps, ambitions — and guide them with the intelligence of a real career coach. Not a chatbot. Not a search bar. A platform that thinks. The name says it all: the only metric that matters.

What it does

Hired. is an AI-powered job search and career development platform that analyses your CV, matches you to relevant job opportunities, and guides you through your career journey with Sara — an interactive 3D AI avatar that acts as your personal career coach. Upload your CV, and Hired. instantly extracts your skills, identifies gaps, ranks job matches by relevance, and delivers personalised recommendations — all in one place. Sara walks you through every step, making the experience feel less like a search engine and more like a conversation with someone who actually knows your profile.

How we built it

We built Hired. across three layers: Frontend — React, with a 3D avatar engine powering Sara's real-time interactions Backend — FastAPI (Python) for all AI/ML inference, and Node.js for authentication, sessions, and job data AI Core — NLP pipelines for CV parsing and skill extraction, semantic similarity for job matching using sentence embeddings: The recommendation engine combines semantic match, experience alignment, and skill coverage into a single relevance score per candidate-job pair.

Challenges we ran into

CV parsing variability — CVs come in every format. Building a robust extractor that handles multi-column PDFs, tables, and inconsistent layouts required multiple preprocessing iterations. Cold start problem — New users have no history. We solved this by weighting CV content heavily at first, then gradually shifting toward behavioural signals as interaction data grows. Sara's latency — Connecting a live 3D avatar to real-time AI responses required streaming output so Sara begins responding before the full model reply is ready. Scope management — Balancing ambition with a student deadline forced us to make hard architectural decisions early and prioritise ruthlessly.

Accomplishments that we're proud of

Built a fully functional AI-powered career platform from scratch as a final-year project Designed and integrated Sara, a 3D interactive AI avatar — something rarely seen in student projects Delivered a working CV analysis + job matching pipeline using real NLP and deep learning techniques Combined a dual-backend architecture (FastAPI + Node.js) cleanly, keeping AI inference decoupled and scalable Produced a platform that solves a real, widespread problem — not just a demo

What we learned

NLP for unstructured text (CVs) is far harder than keyword matching — context is everything Great AI products hide their complexity behind intuitive UX Decoupled microservice-style backends require strict API contracts from day one How to defend architectural decisions clearly to supervisors and a jury That the best learning happens when the stakes are real

What's next for Hired

Recruiter dashboard — letting companies post roles and receive ranked, pre-screened candidates Interview preparation mode — Sara coaches users with role-specific mock questions Multilingual support — Arabic and French CV parsing for the Moroccan and MENA job market Mobile app — bringing the full experience to iOS and Android Analytics layer — giving users visibility into their application funnel and skill progression over time

Built With

Share this project:

Updates