JobPilot: An AI-Powered Job Search Workflow Orchestration Platform

JobPilot is an open-source AI-powered job search management platform designed to address the inefficiencies and missed opportunities caused by scattered information and frequent tool switching during the job search process. It integrates kanban workflow, personal search engine, Gmail automation, and AI task assistant to help users manage the entire job search process systematically. This article will introduce JobPilot from aspects such as background, core features, architecture, and engineering highlights.

Job search is a complex multi-stage process involving job discovery, application tracking, interview preparation, follow-up, and other links. Job seekers often need to switch between multiple tools such as recruitment websites, email clients, calendars, and notes. Scattered information not only reduces efficiency but also easily leads to missed opportunities. JobPilot is designed to solve this pain point.

JobPilot focuses on the goal of "running structured job search activities with less coordination cost" and provides four core functions:

1. Kanban-style Application Tracking: A structured kanban that supports moving positions between different stages (e.g., interested, applied, etc.) and includes task metadata;
2. Personal Search Engine: Aggregates multi-source job information, and performs intelligent sorting through crawlers, standardization, and scoring filters;
3. Gmail Integration & Signal Detection: Automatically identifies key emails such as interview invitations and rejection letters, and maps them to job cards to update status;
4. AI Task Assistant: An intelligent agent embedded in the workflow that parses job descriptions, generates cover letters, and provides personalized suggestions.

JobPilot adopts a multi-component system with clear responsibilities:

• Main Application (SvelteKit): Provides UI and routing, handles authentication, kanban, Gmail connection, etc.;
• Application Backend (Convex): Handles user state, task persistence, and workflow operations;
• Personal Search Backend (Python + FastAPI): Responsible for crawlers, standardization, sorting, and search orchestration;
• Data Persistence: Convex stores application state, while Supabase/Postgres stores search archives, etc.;
• Evaluation Layer: Supports measurable iterative improvements.

JobPilot demonstrates mature engineering thinking in AI integration:

1. Hybrid AI Logic: Combines deterministic rules, structured persistence, domain matching, and AI assistance to balance reliability and flexibility;
2. Reliable Task Orchestration: Includes cleanup processes, retry logic, stuck recovery, etc., with a focus on failure modes;
3. Evaluation-Driven Development: Tracks quality improvements through datasets and scorers, which is data-driven rather than based on subjective intuition.

JobPilot is suitable for the following users:

• Serious job seekers: Track the progress of multiple applications;
• Professionals who change jobs frequently: Maintain active job search channels;
• New entrants to the workforce: Build good job search habits;
• Recruitment consultants: Help clients manage the job search process.

As an open-source project, JobPilot is not only a practical tool but also a production-level AI application architecture example. Its architecture documents, evaluation modules, and deployment configurations reflect professional engineering practices, providing valuable references for developers learning complex AI systems and having important educational value.

JobPilot represents a mature paradigm for AI application development: using AI as an intelligent component combined with deterministic logic, structured data, and professional workflows, rather than a one-size-fits-all solution. Whether you are a job seeker or an AI system developer, JobPilot is an open-source project worth paying attention to.