JaJa: Job Agency Harness (github.com)

JaJa, a new job-search agency utilizing Claude Code, automates the process of finding and applying for jobs while putting the user's approval and data security at the forefront. This system employs a team of specialized AI agents that discover job postings, evaluate fit based on the user's qualifications, draft tailored messages, and only send communications after receiving explicit approval from the user. The architecture ensures that all personal data remains secure within the user’s own data folder, emphasizing a structured and auditable process over automation without oversight. The significance of JaJa lies in its blend of AI capabilities with user control, providing a sophisticated alternative to traditional job application systems. Each action taken by the AI is closely monitored, ensuring that claims made in applications are backed by user-sourced evidence. This not only enhances the applicant's presentation but fosters a sense of ownership and accountability throughout the job search process. By supporting platforms like Gmail and Calendar while keeping a strict separation of personal data, JaJa addresses critical concerns around privacy and data management, making it a compelling tool for engineers and other technical job seekers who value meticulousness in their application journeys.