Build a Small Agent or RAG Tool with the Mistral API
Design and Implement a Small LLM Agent or RAG Tool
You are tasked with designing and implementing a minimal, reliable system that leverages a Large Language Model (LLM) API to build either a small agent or a Retrieval-Augmented Generation (RAG) tool.
The system should be capable of:
- Calling a generic chat-completions style LLM API.
- Retrieving relevant context when necessary.
- Returning grounded answers based on the retrieved context and LLM output.
Constraints and Assumptions
- The exact LLM API surface is not specified; design against a generic chat-completions style API.
- The tool should be small enough to be built and demonstrated in a live coding interview setting.
- Secrets (e.g., API tokens) must be handled securely and not logged.
- AI coding assistants are permitted during the interview.
AI interview prep
Core dilemma
Balancing the simplicity required for a live coding exercise with the inherent complexities of LLM integration, context retrieval, and ensuring 'grounded' and reliable answers. The core choice between a RAG system and a more dynamic agent also presents a dilemma.
3 tradeoffs and 3 stress tests locked
Interview prep
Unlock the interviewer X-Ray rubric, tradeoff matrix, stress tests, clarifying questions, and key concepts.
Build a Small Agent or RAG Tool with the Mistral API
Design and Implement a Small LLM Agent or RAG Tool
You are tasked with designing and implementing a minimal, reliable system that leverages a Large Language Model (LLM) API to build either a small agent or a Retrieval-Augmented Generation (RAG) tool.
The system should be capable of:
- Calling a generic chat-completions style LLM API.
- Retrieving relevant context when necessary.
- Returning grounded answers based on the retrieved context and LLM output.
Constraints and Assumptions
- The exact LLM API surface is not specified; design against a generic chat-completions style API.
- The tool should be small enough to be built and demonstrated in a live coding interview setting.
- Secrets (e.g., API tokens) must be handled securely and not logged.
- AI coding assistants are permitted during the interview.