HireIQ

HireIQ turns any job description into a custom, cheat-proof assessment instantly. With strict tab-monitoring and AWS DynamoDB, we automate the hardest part of technical recruiting.

Comment

Inspiration

We noticed a massive disconnect in technical recruiting. Companies use generic, one-size-fits-all coding tests for highly specific roles, and candidates easily game the system by switching tabs to use ChatGPT. We wanted to build a platform that solves both problems at once: creating bespoke, role-specific assessments instantly using AI, while ensuring absolute integrity through strict browser monitoring.

What it does

HireIQ is an AI-powered B2B hiring dashboard. A recruiter simply pastes a job description, and the platform instantly generates a custom technical assessment containing multiple-choice, text, and live coding questions.

Candidates take the exam in a strict, monitored environment featuring an industry-standard Monaco code editor. We implemented a zero-tolerance anti-cheat system: if the candidate switches tabs or minimizes the window, the exam instantly auto-fails and locks. All candidate profiles, test data, and final scores are securely written to AWS.

How we built it

We built the frontend and API architecture using Next.js 14 (App Router) and deployed it on Vercel for edge-optimized performance. The UI was crafted using Tailwind CSS, shadcn/ui, and @monaco-editor/react to give candidates a real IDE experience.

For the core AI engine, we leveraged the Vercel AI SDK integrated with the DeepSeek API to dynamically generate and structure the test JSON based on the recruiter's prompt.

For the backend, we utilized AWS DynamoDB via the AWS SDK v3. By using Next.js Server Actions, we securely handle the writing and reading of candidate profiles, test schemas, and exam scores directly to AWS without exposing credentials to the client.

Challenges we ran into

Getting the LLM to consistently return a strict, parsable JSON array of questions with varying types (coding vs. multiple choice) was challenging. We overcame this by heavily utilizing the structured output capabilities of the Vercel AI SDK. Additionally, managing the state of the Monaco Editor while simultaneously tracking the anti-cheat visibilitychange DOM events required precise React hook synchronization to avoid false-positive cheating flags.

Accomplishments that we're proud of

We are incredibly proud of the seamless integration between the Vercel frontend and the AWS backend. Architecting a frictionless flow where an AI generates a test via Vercel's edge, a user takes it in a monitored React environment, and the results instantly persist to a NoSQL DynamoDB table makes this feel like a production-ready enterprise product.

What we learned

We leveled up our cloud architecture skills, specifically learning how to securely integrate AWS SDK v3 (DynamoDB) directly into Next.js Server Actions. We also gained massive insights into managing streaming UI states and structured JSON generation using the Vercel AI SDK.

What's next for HireIQ

Next, we plan to implement AWS S3 for secure CV uploads and automated parsing. We also want to build an AI auto-grader that evaluates the candidate's custom code execution, and integrate NextAuth for enterprise-grade recruiter SSO.

Built With

Share this project:

Updates