How pankaj-22/agenticai_poc works
Project type and purpose
• This is a CLI‐style AI proof-of-concept (PoC) tool: “Resume ATS Score Checker using CreAI framework.”
• It ingests a resume, orchestrates a set of AI agents/tasks defined in YAML, calls the Google Gemini API, and generates a scoring summary (output/summary.md).
File structure
• Top level
– LICENSE, README.md, pyproject.toml, requirment (dependency list)
– googleGemini/ (helper module containing an API test for Gemini)
– output/ (where the final summary.md lands)
• src/resume/
– main.py (entry point)
– crew.py (orchestrator of AI agents)
– agents.py (agent definitions)
– tasks.py (task/workflow definitions)
• No separate “frontend” – this is a single process CLI app.
README for additional insights
• Relies on two YAMLs (agents.yaml, tasks.yaml) to configure agent roles and workflows
• Uses a .env file for keys (not shown but implied)
• Core framework is CreAI; external dependency is Google Gemini API
Detail without overthinking
• Module as its own component
• Configuration (agents.yaml, tasks.yaml, .env) feeding into the orchestrator
• Final output file path as a sink