Inspiration
Finding a job is often more about fitting in than just having the right skills. We realized that most hiring platforms focus on resumes, not on personality, culture fit, or how long someone will actually stay and thrive in a role. That’s where Brisa Jobs comes in — our mission is to help companies find people who not only can do the job, but want to stay and grow in it.
What it does
Brisa Jobs is an AI agent that helps companies and candidates find high-retention matches based on cultural alignment and peer-validated soft skills. It features:
A Cultural Fit Matcher that ranks candidates based on their alignment with a company’s values and environment.
An Interview Coach that simulates questions and gives feedback on tone, clarity, and professionalism.
A CV Optimizer that tailors resumes to be more attractive to recruiters and ATS systems.
A Referral Validation System where each job seeker is peer-reviewed by at least three people to confirm their character and soft skills.
How we built it
We built Brisa using Python and Flask for the backend and Jinja2 templates for a simple, fast-rendering frontend. The structure follows a clean separation of concerns using models/, routes/, and services/ folders. For AI functionality, we leveraged LLMs (like GPT-4) through prompt engineering to handle candidate analysis, CV enhancement, and interview coaching. The peer validation and ranking logic is handled with a combination of rule-based checks and LLM evaluation prompts.
Challenges we ran into
Balancing LLM creativity with consistency: ensuring the model responses stayed reliable without drifting from the prompt's intention.
Building a framework for peer validation that felt authentic and scalable.
Creating a system that feels personal, not robotic — we wanted users to feel like they were talking to a real coach or guide, not just filling out forms.
Accomplishments that we're proud of
Designed a full AI-powered pipeline from job seeker entry to employer matching.
Developed a unique peer validation system that filters candidates based on human input + AI reasoning.
Built a usable MVP that demonstrates real-time AI interaction with practical, human-like feedback.
What we learned
Prompt engineering is an art — small changes can lead to radically different outputs.
Users care deeply about belonging, not just employment. Our tests showed that emotional resonance matters.
Building AI tools is not just about automation — it’s about augmenting human connection and trust.
What's next for Brisa Jobs
Move the frontend from Jinja2 to React to allow real-time interactivity and scalability.
Introduce a dashboard for employers to visualize candidate fit over time.
Train a lightweight custom model to evaluate tone, empathy, and engagement in candidate responses.
Explore partnerships with DEI-focused organizations to promote equitable, culture-first hiring.
Log in or sign up for Devpost to join the conversation.