Hirely

AI interview analysis platform that tells you exactly why you failed AI first-round interviews and helps you improve to get hired, solving the #1 reason candidates get rejected.

Comment

Inspiration

A lot of junior undergrad students at my college came to me asking for help with their online interviews, and I could never quite figure out why these bright students were consistently failing first-round interviews. When I investigated the actual interview platforms, I discovered there's absolutely no transparency into what's happening in the background or how candidates are being analyzed. These platforms operate on a rigid set of metrics given to a machine that analyzes you, which is nowhere close to actual peak human recruiters who are much sharper and can see potential beyond checkboxes.

This created a frustrating black box where aspiring professionals were being rejected without understanding their shortcomings. I realized we needed to use the same technology and apply the same metrics to practice, rallying against this exact cause so students can cross first-round interviews and get to the human recruiters who can truly evaluate their potential.

What it does

Hirely is an AI-powered platform that demystifies the AI interview process by scraping real job postings from LinkedIn using BrightData MCP to understand exactly what skills and requirements companies are looking for. Our Groq-powered AI then analyzes these requirements and generates personalized mock interview questions that mirror the actual AI interview experience.

Crucially, Hirely doesn't just ask questions - it analyzes responses using the same metrics and patterns that AI interview platforms use, providing detailed feedback on communication, technical accuracy, and soft skills. We tell candidates exactly why they might be failing, allowing them to refine their approach and confidently pass the initial AI gatekeepers to reach human recruiters.

How we built it

We built Hirely as a robust, asynchronous web service using FastAPI for the backend, ensuring high performance and scalability. The core scraping functionality is powered by BrightData MCP, which allows us to reliably extract up-to-date job market data from LinkedIn. This data is then fed into our AI analysis pipeline, utilizing Groq's powerful language models to process job descriptions, extract key skills, and generate contextually relevant interview questions.

We integrated this with our existing Crawl4AI service, enhancing its capabilities to include job market insights. Pydantic schemas ensure data integrity across our API endpoints, and a clean, modular service architecture allows for easy expansion and maintenance. The frontend is built with React and TypeScript for a seamless user experience.

Challenges we ran into

The most significant challenge was implementing real-time computer vision analysis using OpenCV and MediaPipe to analyze candidate behavior during mock interviews. Getting accurate facial landmark detection, emotion recognition, and posture analysis working seamlessly together required extensive fine-tuning and optimization. We had to handle various lighting conditions, camera angles, and device capabilities while maintaining real-time performance.

Integrating the computer vision pipeline with our AI analysis system was another major hurdle, as we needed to correlate visual cues with verbal responses to provide comprehensive feedback. Ensuring the system worked consistently across different devices and browsers while maintaining low latency was particularly challenging.

Accomplishments that we're proud of

We are incredibly proud of creating a system that directly addresses a critical, often overlooked, pain point in the modern job market. Hirely empowers junior candidates by providing unprecedented transparency into the AI interview process. Our ability to scrape real-time job data and translate it into personalized, actionable interview preparation is a significant achievement.

We've successfully built a platform that not only identifies what skills are needed but also how to articulate them effectively in an AI-driven environment, giving candidates the confidence and tools to succeed where they previously failed. The fact that we're helping students from our own college community makes this even more meaningful.

What we learned

Through building Hirely, we gained deep insights into the evolving landscape of AI in recruitment and the critical need for tools that help candidates navigate it. We learned the importance of precise data extraction for effective AI analysis and the nuances of generating truly personalized and impactful feedback.

We also reinforced the value of a modular, service-oriented architecture for integrating diverse technologies like web scraping, AI, and web frameworks. Most importantly, we learned that by understanding the "rules" of the machine, we can empower humans to outperform them and reach the human recruiters who can truly evaluate their potential.

What's next for Hirely

For Hirely, the future involves expanding our AI analysis to include more sophisticated behavioral and situational interview simulations. We plan to integrate with more job platforms and potentially offer company-specific AI interview insights. Further enhancements will include real-time audio/video analysis during mock interviews to provide even more granular feedback on non-verbal cues and communication style.

Our ultimate goal is to become the go-to platform for anyone looking to master AI first-round interviews and confidently secure their dream job by getting past the machines and reaching the human recruiters who can see their true potential.

Built With

Share this project:

Updates