About the project
The project started as a tool for reviewing my mentees' interviews. It has grown into an ecosystem for preparing AI engineers and running mentorship. Participants get learning, practice, and interview assistance; I get CRM tools, intake support, mentoring history, and analytics. These are separate services, not a single application.
Learning materials cover four phases: production Python, LLM fundamentals, RAG, and agents. Theory connects to hands-on projects, real take-home assignments, and their solutions.
The real interview library shows what companies ask and how different rounds work. Learners can search questions by topic, study reviews, and jump to the relevant moment in a recording.
Candidates can upload their own interviews for AI-assisted review. The system transcribes recordings, extracts questions, evaluates answers, and highlights topics and communication skills to improve. A mentor reviews and expands the recommendations; anonymized reviews can be published to the library.
A custom MCP server connects the platform to ChatGPT or Claude. Learners can search materials and questions, explore topics, and practice interviews through their usual AI assistant. With separate consent, it can also use their own interview feedback and individual meeting summaries, never someone else's private context.
Orbyte is a separate desktop overlay assistant for AI engineering interviews. It uses the candidate's context and the ongoing conversation, helps with questions and on-screen tasks, and displays hints during the interview.
The internal CRM tracks people and mentorship: profiles, statuses, agreements, payments, access, notes, and meeting summaries. It keeps the mentoring history together so the next step can build on previous work rather than scattered messages and documents.
Telegram bots support intake by collecting a person's experience and goals, answering from a curated knowledge base, and escalating questions that need my personal decision. Agreement terms and important decisions remain my responsibility.
Analytics from published interviews reveal recurring questions, common weak spots, and topics worth studying. Separately, website analytics show how people find and use the site. Both inform the learning product and the mentorship business.