SWIMR

Home Page
Analyser
Candidate Profile
Discover Page
Discover Page (Dark)
Roles
Agent Financial Observability
Per-Unit Economics and ROI Projection Calculator
Transaction Log and Projected Cumulative Value
Personalised Acceptance Email

Inspiration

We started with a simple realization that the recruitment industry is deeply flawed. Reading and filtering resumes is a genuinely exhausting task that drains company resources and naturally introduces human bias. We know this firsthand because we have family members working in the HR sector, and we have heard endless frustrations about the hours wasted manually scanning resumes that often do not tell the whole story.

Companies are spending tens of thousands of dollars on agency fees just to get a stack of papers that an engineering manager still has to spend hours reviewing. We wanted to build something that actually solves this problem, something that saves companies thousands of hours and evaluates candidates based on their actual work rather than just what they claim on a PDF.

What it does

SWIMR is an AI technical recruiter that handles the heavy lifting of the hiring process. We built a complete set of tools for HR teams to use easily.

For starters, recruiters can drop massive files of resumes into our system all at once, and our AI instantly reads, organizes, and analyzes them against specific job descriptions. Instead of a messy email inbox, recruiters get a clean leaderboard where candidates are ranked side by side based on a fair AI match score. This shows exactly which skills they have and what they might be missing.

We also wanted to actively find talent rather than just wait for it. By typing in a job description, our Discover page acts as a talent scout. It searches the GitHub API to find developers based on their location, languages, and actual coding history. Finally, we added a dashboard that tracks exactly how much time and money the agent is saving the company in real time by comparing our computing costs to traditional human labor hours.

How we built it

We used Lovable to quickly build our frontend, constantly iterating to create a clean and professional interface in React. The actual brain of the operation is powered by Claude, which acts as our expert engineering manager to score the candidates fairly. For our data pipeline, we connected securely to the GitHub API, and we orchestrated all the complex background logic through secure edge functions.

Challenges we ran into

Building a fully functional AI agent during a hackathon really tested our patience and teamwork. At first, we faced some major hurdles figuring out how to legally and safely source candidate data without violating strict terms of service on platforms like LinkedIn.

We also went out of our way to ask for critical feedback while we were building. Hearing brutally honest critiques about our early versions was tough to swallow, but it was exactly what we needed. It forced us to rethink our approach and become much more ambitious, pushing us to transition from just making a simple resume reader into building a proactive and legally compliant sourcing agent.

Accomplishments that we're proud of

We are incredibly proud of our teamwork and how quickly we adapted to harsh feedback under a lot of pressure. On the technical side, we are really proud of building a platform that is highly scalable and strictly GDPR compliant. By switching our focus to the official GitHub API for our Discover page, we completely avoided the legal headaches of scraping closed websites, making sure our tool is actually ready for the real world.

What we learned

We learned that your first idea is rarely your best idea. Learning to take critical feedback on the chin and actually using it to change direction made our project so much better than we originally planned. We also learned a ton about how to connect complex AI data pipelines, handle API limits securely, and blend smart backend logic into a fast and responsive user interface.

What's next for SWIMR

Right now, our sourcing engine is highly specialized for the tech sector because we use GitHub to find objective proof of a candidate's work. Moving forward, we want to scale SWIMR into a more generalized AI recruiter. We plan to connect it with other industry databases and professional portfolios, allowing the agent to scour different sources to find, rank, and verify great candidates for any role, whether that is in design or finance, all while keeping our strict data privacy standards intact. With scaling, we can also hope to integrate SWIMR into existing ATS systems used by corporations, like Workday, so SWIMR can function as a plugin.