Python Backend Developer for AI Demo
Budget: $1,500 – $3,000 USD
Required Coding Knowledge
You will be expected to be a Python backend developer with significant knowledge of modern, asynchronous programming and an understanding of distributed systems.
• Python (Advanced): You must be highly proficient in Python 3.10+. This includes advanced knowledge of asyncio for handling asynchronous operations, context managers for resource management, and a deep understanding of Python's ecosystem, including Pydantic for data validation, FastAPI for building APIs, and Jinja2 for templating.
• Infrastructure & DevOps: You should have a working knowledge of containerization with Docker and understand how to manage a local Redis instance for caching and state management. You will also need a basic understanding of SQL and ORM concepts (e.g., SQLAlchemy and sqlite+aiosqlite) to work with the platform's mock databases.
• AI/ML Concepts: You don't need to be an ML expert, but you must have a solid understanding of how to interact with LLMs via API calls. This includes an understanding of concepts like prompt engineering, streaming responses, and handling API-specific errors.
• Security & Best Practices: You must have a strong grasp of security fundamentals. This includes understanding how to handle secrets securely, prevent path traversal attacks, and redact sensitive information from logs. You'll need to know where to enable or disable production-only security features for the demo.
Your Mission: The Path to a Compelling Demo
• Create the "Happy Path": Focus on one compelling, end-to-end flow. For example, show the clarifier taking a user's ambiguous request, turning it into a structured plan, sending it to the codegen_agent, having the critique_agent find a minor issue, and then seeing the refactor_agent apply the fix.
• Simplify the Infrastructure: You will bypass the platform's extensive, production-ready infrastructure. This means using a local file backend for auditing and a mock LLM provider for intelligence, as suggested by the project's own local_provider.py and dlt_simple_clients.py files. You will have to create a simplified version of the demo, as a fully-featured version is not in the code.
• Debugging the Un-run Code: Much of the code, particularly in the most complex parts like dlt_evm_clients.py and gremlin_chaos_plugin.py, has likely never been run end-to-end. When you encounter a bug, you won't just be debugging a typo; you'll be debugging a logical error in a complex, asynchronous, and potentially un-tested code path.
• The High-Friction Onboarding Process: The platform's onboarding process is a gauntlet of dependencies and configurations. You'll need to meticulously unify configurations from scattered files, set environment variables, and manage a sprawling list of dependencies.
• Trust the Vision: Your job is not to build this platform. That’s been done. It's to be the brilliant integrator who shows that all the pieces of this complex machine can work in harmony. The platform's architecture is a testament to what AI can build, and your demo will be a testament to what a skilled human can do with an AI-generated product. By creating a compelling, end-to-end demonstration, you will have successfully proven why a larger team should be brought on to take it to the next level.
You will be expected to be a Python backend developer with significant knowledge of modern, asynchronous programming and an understanding of distributed systems.
• Python (Advanced): You must be highly proficient in Python 3.10+. This includes advanced knowledge of asyncio for handling asynchronous operations, context managers for resource management, and a deep understanding of Python's ecosystem, including Pydantic for data validation, FastAPI for building APIs, and Jinja2 for templating.
• Infrastructure & DevOps: You should have a working knowledge of containerization with Docker and understand how to manage a local Redis instance for caching and state management. You will also need a basic understanding of SQL and ORM concepts (e.g., SQLAlchemy and sqlite+aiosqlite) to work with the platform's mock databases.
• AI/ML Concepts: You don't need to be an ML expert, but you must have a solid understanding of how to interact with LLMs via API calls. This includes an understanding of concepts like prompt engineering, streaming responses, and handling API-specific errors.
• Security & Best Practices: You must have a strong grasp of security fundamentals. This includes understanding how to handle secrets securely, prevent path traversal attacks, and redact sensitive information from logs. You'll need to know where to enable or disable production-only security features for the demo.
Your Mission: The Path to a Compelling Demo
• Create the "Happy Path": Focus on one compelling, end-to-end flow. For example, show the clarifier taking a user's ambiguous request, turning it into a structured plan, sending it to the codegen_agent, having the critique_agent find a minor issue, and then seeing the refactor_agent apply the fix.
• Simplify the Infrastructure: You will bypass the platform's extensive, production-ready infrastructure. This means using a local file backend for auditing and a mock LLM provider for intelligence, as suggested by the project's own local_provider.py and dlt_simple_clients.py files. You will have to create a simplified version of the demo, as a fully-featured version is not in the code.
• Debugging the Un-run Code: Much of the code, particularly in the most complex parts like dlt_evm_clients.py and gremlin_chaos_plugin.py, has likely never been run end-to-end. When you encounter a bug, you won't just be debugging a typo; you'll be debugging a logical error in a complex, asynchronous, and potentially un-tested code path.
• The High-Friction Onboarding Process: The platform's onboarding process is a gauntlet of dependencies and configurations. You'll need to meticulously unify configurations from scattered files, set environment variables, and manage a sprawling list of dependencies.
• Trust the Vision: Your job is not to build this platform. That’s been done. It's to be the brilliant integrator who shows that all the pieces of this complex machine can work in harmony. The platform's architecture is a testament to what AI can build, and your demo will be a testament to what a skilled human can do with an AI-generated product. By creating a compelling, end-to-end demonstration, you will have successfully proven why a larger team should be brought on to take it to the next level.