This is a submission for the Hacktoberfest Open-Source AI Challenge Week 2: Side Hustle
What I Built
I built HireLens AI, an AI-powered recruitment and applicant tracking platform designed for small businesses, startups, and recruitment agencies.
Hiring is expensive and time-consuming, especially for small teams that cannot afford dedicated recruitment departments or expensive enterprise hiring software.
Recruiters often spend hours:
- Reading and extracting information from resumes.
- Comparing candidates against job requirements.
- Shortlisting applicants.
- Tracking candidates through recruitment stages.
- Making sense of different candidate profiles.
HireLens AI brings these workflows into one recruitment workspace.
Its key capabilities include:
- Resume Intelligence: Extract structured information from uploaded resumes.
- Job Management: Create and manage job openings.
- AI-Assisted Matching: Compare candidate profiles against job requirements.
- Explainable Matching: Understand why a candidate matches a particular role.
- Candidate Comparison: Compare multiple applicants side by side.
- Recruitment Pipeline: Track applicants throughout the hiring process.
- Analytics: Understand recruitment activity and pipeline performance.
The goal is simple: give smaller recruitment teams access to useful hiring intelligence without the complexity and cost of enterprise recruitment platforms.
Demo
Live Application: https://hire-lens-black.vercel.app
The application includes a one-click demo experience.
Visitors can explore the platform as a fictional demo recruiter, Alex Morgan, without creating an account or entering credentials.
The demo includes synthetic candidates, sample job openings, resume workflows, and explainable matching results.
The platform also protects its demo workspace against destructive operations and isolates demo data from other tenants.
Note: The free backend hosting may take a little time to wake up after inactivity.
Code
GitHub Repository: https://github.com/kusalar/HireLens
The project uses a full-stack architecture:
- Frontend: Next.js 14, React and Tailwind CSS.
- Backend: Python, FastAPI and Pydantic.
- Database: PostgreSQL with SQLAlchemy and Alembic.
- Background processing: Celery and Redis.
- AI: Open-weight model(Ollama) integration for resume intelligence and grounded explanations.
- Matching: Deterministic Python-based scoring engine.
How I Built It
I wanted HireLens AI to be more than another chatbot that generates recruitment recommendations.
I designed it around a combination of open-weight AI, conventional backend engineering, and deterministic decision logic.
1. Resume Intelligence
The platform extracts information from PDF and DOCX resumes and converts it into structured candidate information.
The architecture supports open-weight model inference, including local development with Ollama and hosted open-weight inference.
2. Deterministic Candidate Matching
One of the most important architectural decisions was separating candidate scoring from language model generation.
The matching score is calculated using deterministic Python logic.
The LLM does not decide who should be hired, does not autonomously reject candidates, and does not control the final matching score.
Instead, AI assists with grounded explanations that help recruiters understand the matching results.
3. Recruitment Workspace
I built the recruitment workflow around jobs, candidates, applications, resume processing, matching, comparison, and pipeline management.
The backend uses FastAPI, PostgreSQL, SQLAlchemy, and tenant-aware data access to support multiple recruitment organizations.
4. Production Deployment
I deployed the frontend on Vercel, the backend on Render, and the PostgreSQL database on Neon.
I also implemented a publicly accessible one-click demo mode so evaluators can explore the actual application.
The current deployment uses free-tier infrastructure, making it possible to demonstrate the product without maintaining expensive hosting services.
Why Does Open Innovation Matter?
For HireLens AI, open innovation is not just about using a different AI model. It is about having control over the technology that powers the product.
Recruitment platforms process sensitive information. Resumes can contain contact details, employment histories, education records, and other personal information.
Building around open-weight models gives us more control over where inference happens and how models are deployed.
There are several important advantages:
1. Model Independence
I do not want the entire product to depend on a single proprietary AI provider.
Open-weight models give us the flexibility to experiment with different models, change inference providers, and adapt the system as the product grows.
2. Cost Control
A recruitment SaaS for small businesses needs sustainable economics.
Depending entirely on expensive proprietary inference APIs could make the product difficult to operate at scale.
Open-weight inference gives us the option to explore local and self-hosted deployments and better control inference costs.
3. Data Control
Local inference can provide an additional privacy option for organizations that do not want resume content sent to external AI providers.
It also creates opportunities for future private deployments.
4. Transparent and Controlled Decisions
I deliberately kept candidate scoring deterministic rather than allowing an LLM to make hiring decisions.
This makes matching more reproducible and easier to test, while AI-generated explanations help recruiters understand the results.
For me, open innovation means being able to inspect, adapt, replace, and control important parts of the technology instead of building the entire business around someone else's API.
Built with the idea that powerful recruitment intelligence should be accessible to smaller teams, not just enterprises.
Thanks for checking out HireLens AI!
Top comments (0)