JobWizard: AI-Powered Job Application Automation

bd-mcp-demo-desktop.mp4

Overview

JobWizard is an AI-driven web application that revolutionizes the job search process, drastically reducing time and boosting application success rates. By automating job offer discovery and generating personalized cover letters, JobWizard empowers job seekers to focus on their career goals while the system handles the heavy lifting. Built for the Bright Data Real-Time AI Agents Challenge, JobWizard leverages Bright Data’s MCP server to seamlessly discover, access, extract, and interact with real-time web data, delivering tailored job matches and compelling cover letters with unmatched precision.

Problem Solved

Job searching is time-consuming and often discouraging, requiring hours of manual effort to find relevant opportunities and craft tailored applications. JobWizard addresses this by automating the discovery of job offers and generating personalized cover letters using real-time company and user experience data. Access to reliable, current web data ensures accurate job matches and high-quality cover letters, significantly improving application success rates.

Meeting the Bright Data Challenge Criteria

JobWizard leverages Bright Data’s MCP server to power a multi-agent system that automates job searches and cover letter generation, fulfilling three out of the four key criteria of the Bright Data Real-Time AI Agents Challenge: Discover, Access and Extract. By integrating reliable, real-time web data, JobWizard drastically reduces job search time and improves application success rates, addressing a critical real-world problem.

• Discover: Find Relevant Content Across the Open Web
  JobWizard’s agent-offers-finder uses Bright Data’s search_engine tool to query job boards and platforms like LinkedIn, BuiltIn, and RemoteRocketship in real time. For example, a user seeking “Data engineering jobs in France” triggers precise searches (e.g., “Data engineering jobs in France Python SQL AWS”) to identify relevant job listings instantly. This ensures the system discovers up-to-date opportunities tailored to the user’s resume, enhancing AI performance by providing a broad, current dataset for matching.

• Access: Navigate Complex or Protected Websites
  Accessing dynamic job boards and company websites is challenging due to rate limits and IP blocks. Bright Data’s MCP server simplifies this by enabling seamless navigation of complex, JavaScript-rendered pages. The scrape_as_markdown tool allows JobWizard to access job listings (e.g., Devoteam’s career page at https://jobs.smartrecruiters.com/devoteam/744000042826005) without disruptions, ensuring reliable data retrieval for downstream processing.

• Extract: Pull Structured, Real-Time Data at Scale
  JobWizard extracts structured data from diverse sources using Bright Data’s scrape_as_markdown and web_data_linkedin_company_profile tools. The agent-offer-enricher pulls detailed job and company insights, such as role requirements and company culture from LinkedIn profiles and career pages. This structured data (e.g., job descriptions, salary ranges) is critical for generating accurate, personalized cover letters, improving AI performance by grounding outputs in high-quality, real-time information.

By integrating Bright Data’s MCP server tools with a custom TypeScript multi-agent framework, JobWizard overcomes the challenge of scraping reliable data without rate limits or IP blocks. This ensures real-time, accurate data flows into the AI pipeline, enabling precise job matching and compelling cover letters that boost application success rates.

How It Works

JobWizard’s multi-agent system orchestrates a seamless workflow, powered by Bright Data’s MCP server and a custom TypeScript framework. Here’s how it transforms your job search:

1. Upload Your Resume: Upload a PDF resume, and our AI parser extracts skills, experience, and education into structured data using advanced vision LLMs.
2. Discover Relevant Jobs: JobWizard searches the web in real time for job offers tailored to your profile (e.g., “Data Engineer roles in France”). Bright Data’s MCP server enables discovery and access to platforms like LinkedIn, BuiltIn, and RemoteRocketship.
3. Select Your Top Picks: Review a curated list of job offers and choose roles that excite you, such as an “Azure Data Engineer” position at Devoteam.
4. Enrich Job Insights: For each selected job, JobWizard pulls detailed insights (e.g., company culture, salary ranges) from sources like company websites and LinkedIn profiles using Bright Data’s scraping tools.
5. Get Tailored Cover Letters: Receive personalized cover letters crafted to align your skills with each job’s requirements, ready to apply in minutes.

Demo

Try JobWizard live at https://brightdata-mcp.bitswired.com

Testing Credentials:

• Email: noah@brightdata.com
• Password: Provided by email to Noah to avoid abuse

Usage Notes:

As a proof-of-concept built under time constraints, JobWizard focuses on demonstrating core functionality rather than handling all corner cases or scalability. To ensure a smooth experience, I recommend selecting 1 or 2 job offers when prompted to choose roles, as this keeps the process quick. The full workflow (resume parsing, job discovery, data enrichment, and cover letter generation) involves extensive scraping with our multi-agent system and Bright Data’s MCP tools, so it may take a few minutes to complete.

Mobile and Desktop Compatibility:

JobWizard works seamlessly on both mobile and desktop, but maintaining the WebSocket connection is critical for the process to complete uninterrupted. On desktop, simply avoid refreshing the page. On mobile, please don’t exit the app or let the phone lock, as this may break the WebSocket connection. Reconnecting after unlocking can disrupt the process. This is a straightforward fix that I can implement, but it wasn’t prioritized for the demo as it adds limited value for showcasing the core functionality.

Repository: https://github.com/bitswired/jobwizard.

Here’s JobWizard in action:

https://pub-2841a52b911a4219930a8f0eeb5a5550.r2.dev/bd-mcp-demo-desktop.mp4: Desktop Demo: Watch JobWizard find “Remote Jobs in AI” and generate a tailored cover letter.

https://pub-2841a52b911a4219930a8f0eeb5a5550.r2.dev/mobile-demo.mp4: Mobile Demo: Watch JobWizard find “AI Engineers jobs in Geneva” and generate a tailored cover letter.

The demo showcases the full workflow: upload a resume, discover job offers (e.g., “Azure Data Engineer at Devoteam”), select roles, enrich data, and receive a personalized cover letter, all powered by Bright Data’s real-time web data.

Technical Implementation

JobWizard is built with a modern, scalable tech stack, integrating Bright Data’s MCP server and a custom TypeScript multi-agent framework:

• Bun Monorepo: Houses API (ElysiaJS), frontend with Tanstack Router/Query), and a custom agent package, all in TypeScript.
• Custom Multi-Agent Framework: A TypeScript-based system enabling complex, hierarchical workflows. The agent-router dispatches tasks to specialized agents (agent-offers-finder, agent-offer-enricher, agent-cover-letter-writer) and tools, supporting arbitrary nesting for dynamic task orchestration.
• Bright Data MCP Server: Powers real-time web data tasks:
  • search_engine: Queries job boards (e.g., “Data engineering jobs in France Python SQL AWS”).
  • scrape_as_markdown: Extracts structured data from dynamic pages (e.g., Devoteam’s career page).
  • web_data_linkedin_company_profile: Enriches job offers with company insights from LinkedIn.
• WebSocket: Enables real-time interaction between users and agents, displaying progress (e.g., “Parsing resume…”) and user inputs (e.g., job selection) via UI elements triggered by WebSocket events.
• Database: SQLite for lightweight, efficient storage of job data and user profiles.
• Deployment: Docker Compose on Hetzner, proxied by Cloudflare for reliability and performance.

Challenges Overcome

Scraping accurate data from websites like search engines, LinkedIn, and Glassdoor is challenging due to rate limits and IP blocks when using backend automated tools. Bright Data’s MCP server simplified this by providing robust scraping infrastructure, enabling seamless access to dynamic, JavaScript-rendered pages. Tools like search_engine, scrape_as_markdown, and web_data_linkedin_company_profile allowed JobWizard to reliably discover job listings, extract structured data, and enrich offers with company insights, ensuring high-quality, real-time results without disruptions.

Why JobWizard Stands Out

• Real-World Impact: Saves hours of manual job searching and tailors applications to boost success rates, directly addressing a common pain point for job seekers.
• Real-Time Data: Leverages Bright Data’s MCP server for up-to-date job matches and enriched insights, improving AI performance with reliable web data.
• Technical Innovation: The custom TypeScript multi-agent framework enables complex, nested workflows, making JobWizard scalable and adaptable.
• User Experience: Real-time WebSocket updates and an intuitive UI ensure a seamless, human-in-the-loop interaction.

Future Improvements

• Expand job source coverage
• Enhance cover letter customization with user feedback loops.
• Integrate more advanced LLMs for deeper personalization.
• Generate entirely customized resumes for each job offer

Acknowledgments

JobWizard was built for the Bright Data Real-Time AI Agents Challenge, utilizing their excellent MCP server tools. Special thanks to the Bright Data team for providing $250 in credits and support via noah@brightdata.com.

Contact

For questions or feedback, reach out at jimi@bitswired.com or X for updates.

---

Built with ❤ for the Bright Data Challenge