Inspiration: As a second-year engineering student at Sri Sairam Institute of Technology in Chennai, I was actively searching for internships and job opportunities when I came across something deeply unfair. Candidates with well-crafted but fake resumes were getting shortlisted, while genuinely skilled individuals were being rejected. Recruiters had no reliable way to check if a candidate's resume matched their actual skills or interview performance. I saw friends with real talent get overlooked simply because they didn’t know how to navigate the system. That frustration became my driving force. I asked myself — what if AI could detect the fakes and find the genuine talents? That question led me to create GenuAI Technologies, an AI-powered recruitment intelligence platform that cross-checks candidates across their resume, skills, and interview performance using a Triangle Consistency Engine. GenuAI wasn’t just a hackathon project — it was a solution to a real problem I had personally experienced.
What it does: GenuAI Technologies is an AI-based platform that helps with hiring by finding out if job applicants are real or not. It checks candidates using something called the Triangle Consistency Engine, which looks at three different scores: the resume score from the ATS system, the result of a skills test, and the interview score. If someone has a high resume score but a low skills test score, the system flags them as suspicious. But if all three scores match up, the candidate is considered genuine. The platform offers role-specific multiple choice tests for eight different job roles, allows for video and voice pitch recordings, and uses AI models like Google Gemini and Anthropic Claude for evaluation. It also has strong security features with 11 layers to prevent cheating, detects fake resumes, analyzes personality through five traits, shows skill gaps with a heatmap, tracks the candidate's journey over time, gives a lie detector score, uses voice stress analysis, detects emotions in real time, checks typing patterns, offers an AI-powered mock interview coach, and includes tools to create resumes and cover letters with AI suggestions. Plus, it has a company intelligence dashboard that lets you compare candidates.
How we built it: We built GenuAI entirely on AWS within 8 days during the AI-PrepPulse Hackathon. The frontend was developed using React 18 and TypeScript with Vite, hosted on AWS S3 and delivered via CloudFront CDN. The backend was built using Node.js and Express with TypeScript, running on an EC2 t3.micro instance, managed by PM2 and Nginx, with 20 API routes. The database is PostgreSQL 16.6 on AWS RDS, featuring 8 tables that store users, assessments, questions, test answers, cheat logs, jobs, interviews, and email logs. Five AWS Lambda functions, written in Python 3.12, handle ATS resume scoring, fake resume detection, MCQ skill scoring, triangle consistency calculation, and full AI evaluation using Google Gemini, with Anthropic Claude as a backup option. AWS SES is used for sending transactional emails related to OTP verification, assessment results, hire and reject notifications, and interview scheduling.
Challenges we ran into: The main technical issue was a mismatch between the installed PDF.js version 5.5.207 and the CDN worker version 3.11.174, which led to resume reading failing silently without any error messages. We resolved this by switching to a bundled worker URL. Developing an 11-layer anti-cheat system that prevents tab switching, keyboard shortcuts, copy and paste, right-click actions, screenshots, picture-in-picture, screen capture, dragging, and window blur at the same time while still allowing the interview to function properly was very challenging. AWS Lambda had cold starts taking 3 to 5 seconds on the first run, which impacted the user experience and needed Python code optimization. The EC2 instance's public IP changed every time it was rebooted, causing repeated SSH connection timeouts and requiring updates to the Security Group rules each time. Handling over 800 lines of TypeScript code with complex JSX nesting led to multiple build issues that required Python scripts for safe file modifications.
Accomplishments that we're proud of: We are really proud of winning first place in the AI-PrepPulse Hackathon, which had over 60,000 participants from across the country, as a second-year engineering student. In just eight days, we built more than 128 features on a real AWS production environment with a live URL, which we are very proud of. The Triangle Consistency Engine is a completely original idea we came up with from scratch to solve a real problem in the hiring process. By implementing 11 security layers in the interview room—including real-time motion detection, emotion detection, voice stress analysis, and keystroke dynamics—we made GenuAI one of the most secure candidate evaluation platforms created by a student. The platform is fully live, functional, and open source on GitHub.
What we learned: We studied the full-stack AWS architecture by integrating services like EC2, RDS, S3, CloudFront, Lambda, API Gateway, and SES into a single, smooth platform. We explored AI API integration by using Google Gemini and Anthropic Claude as evaluation tools, along with backup plans for when things don’t work. We focused on security engineering by creating multi-layer anti-cheat systems for proctoring. We practiced using TypeScript at a large scale by handling over 5000 lines of typed code in React and Node.js across six pages and eight backend routes. We developed product thinking by designing features that address real challenges in hiring, not just focusing on building technology. Most importantly, we realized that staying persistent through bugs, timeouts, version issues, and build failures is what makes the difference between a completed product and an unfinished idea.
What's next for GenuAI Technologies: GenuAI is working to become the standard tool for hiring verification in India, replacing the manual screening done by HR teams with AI-driven consistency checks. The next steps involve launching a custom website at genuaitechnology.com, gaining access to AWS SES for sending emails to candidates around the world, integrating real TensorFlow-based face detection for better proctoring, developing a mobile app for candidates on both Android and iOS, adding AI-generated interview questions based on each candidate's resume using the Claude API, setting up a subscription billing system for companies using Razorpay, and expanding to cover more than 50 job roles instead of just the current 8. In the long run, we plan to team up with colleges and placement offices across India to make GenuAI the go-to platform for campus recruitment. We don't just look at resumes. We check if the person is who they say they are.
Log in or sign up for Devpost to join the conversation.