Preparing for job interviews can be stressful—especially when you’re trying to research the interviewer and company right before the big day. I recently built a project to solve exactly that: an AI Interview Agent that prepares a tailored summary of both the company and the interviewer so you walk in ready.

In this post, I’ll walk through the idea, tools used, and how everything comes together to generate the final interview prep summary.

✨ What Does the Interview Agent Do?

Imagine you have an upcoming interview and you’ve got the interviewer’s LinkedIn profile and the company’s website. You feed both into this agent, and it gives you a smart, concise summary including:

  • Insights about the interviewer (role, background, interests)
  • Summary of the company (industry, focus, values)
  • Possible interview questions you might be asked based on both

It’s like having a mini research assistant—powered by AI.

🔧 Tools & Tech Stack

Here’s what powers the Interview Agent:

  • Relevance AI: Used to fetch and analyze the interviewer’s LinkedIn profile.
  • Firecrawl: Crawls the company’s website to extract relevant information.
  • OpenAI (GPT-4): Used to generate summaries and potential interview questions.

📸 Screenshot: Architecture of the AI Interview Agent
Image

⚙️ How It Works — Behind the Scenes

The logic is split into 3 main steps:

1. Crawl the Company Website

Using Firecrawl, the tool extracts content from the company’s official site. This includes their About page, recent news, team info, and product offerings.

📸 Screenshot: Sample Firecrawl output for a company
Image

2. Analyze Interviewer’s Profile

With the LinkedIn URL, we fetch the interviewer’s details using Relevance AI—like job title, experience, and interests. This info helps us gauge their professional background and approach.

📸 Screenshot: Relevance AI fetching a LinkedIn profile
Image

3. Generate AI-Powered Summaries

All extracted data is passed to OpenAI’s LLM, which does three things:

  • Creates a short summary of the company
  • Crafts a summary of the interviewer
  • Based on both, predicts what questions the interviewer may ask

📸 Screenshot: What final tool looks like Image

đź“Ś Use Cases & Future Ideas

This tool is great for:

  • Job seekers preparing for interviews
  • Career coaches helping clients prep
  • Mock interview platforms integrating smart AI helpers

In the future, I’d love to add:

  • PDF export of the summary
  • Integration with calendars to auto-trigger before an interview
  • Voice-read summaries for on-the-go prep

🚀 Final Thoughts

This project was a fun and practical way to explore multi-agent AI systems and real-world applications of LLMs. It also highlights the power of combining structured crawlers, smart data enrichment, and language models.

If you’re interested in the technical breakdown or want to try it out, feel free to connect!

🧠 Personal Thoughts

While OpenAI has been incredibly powerful for generating results, I’m currently exploring self-hosted alternatives to give me more flexibility and control. One of my goals is to run an open-source LLM like DeepSeek on my personal server. This way, I can eliminate external restrictions, keep everything on-prem, and potentially enhance privacy and customization for future versions of this tool.