ACETRAIT

Acetrait is an AI-powered platform designed for the recruitment process for both companies and job seekers. At its core, Acetrait simplifies and automates the task of creating engaging job posts, developing ideal candidate personas, conducting in-depth interviews, and evaluating candidates through AI and LLMs.

Initially, I wanted to build the application with svelte in sveltekit wishing to take advantage of Svelte's nice global state logic, but my love for next.js combined with its amazing developer experience and the latest Server Actions made me go with Next.js. Back into the world of react , with tailwindcss , I started to have fun coding the front end in a matter of days.

I built the server mostly utilising the new Server Actions, with nodejs , and hosted on vercel ofc. Also used edge functions for couple of things in the application, to save money by not running function on the servers as there were long waiting periods for response from some API fetch calls from the LLMs, frequently hitting the Server Execution Limits on Vercel.

What else? I fully committed to typescript . The database is built on supabase , and I was quickly able to set up secure cookie based auth and in no time I had a highly integrated RLS backed postgresql database, just waiting for my server actions to come fetch and feed them. The AI is powered by openai API.

• AI-Driven Interviews

  Personalized, in-depth chat-based interviews that evaluate a candidate's fit for the job position. The AI is trained to ask relevant questions but using a dynamic real time prompting technique.

• Candidate Evaluation and Scoring

  AI assesses candidates across multiple attributes, presenting results in a spider chart. In order to make the AI scoring more reliable, the latest API attributes from OpenAI are used to their full potential.

• Automated Job Post Creation

  Generate captivating job advertisements and ideal candidate personas with just basic job details.

• Job Feed

  Job posts are accessible in a comprehensive public feed for candidates.

Organisation dashboard, where you can create job posts, view applicants, and take AI interview for the organisation, to help the AI understand the organisation better.

Creating job post using AI, just give the basic details and the AI will create the job post for you.

AI grading of the candidate, with the reasoning for the score.

AI grading reasoning, with the attributes and the score.

• Frontend

  Utilizing Next.js 14 and React with Tailwind CSS and Shadcn for a user-friendly interface. The application ensures a modern, fast, and secure user experience.

• Distinct Interfaces for Users

  Features tailored UIs for both candidates and recruiters, equipped with role-specific functionalities.

• Backend and Edge runime

  The backend is built on Node.js, hosted on Vercel, and uses Edge functions for enhanced performance. React Server Components are used for security and efficiency.

• Database and AI Integration

  Supabase is used for the database, ensuring scalability and data integrity. The AI component, powered by OpenAI's API, is fine-tuned for accuracy and relevance.

The final website is still under development, with more innovative features and updates coming to it. But it is already being used by a couple of organisations for testing and trying out the AI features on real interviews. Head over to www.acetrait.com to test it yourself.

This project has led me to better understand AI and LLMs in general, and I had a great developer experience while coding this application. Largely it was thanks to the amazing libraries that simplifies the use of AI for the top LLMs providers like OpenAI, Mixtral etc.