Wannabe vs $200K AI Engineer: Fixing RAG, Agent Tools and Model Choice

Bashiri Smith · Facebook reel · 2026-09-02 · 1:00 · 5,197 views · Open on Facebook

Topics: Retrieval-Augmented Generation (RAG), AI Agents, Tool Use & MCP, LLMOps, Deployment & Monitoring · Level: intermediate

Summary

A short skit comparing a beginner's habits with a senior AI engineer's habits in three areas of production AI. Retrieving more chunks doesn't fix bad RAG answers; measuring retrieval quality and reranking does. Agents work more reliably with a small set of tools, and sending simple tasks to cheaper models protects margins and latency. The last third promotes the creator's AI-engineering community, which you join by commenting "JOIN".

Key points

• Retrieving more chunks doesn't automatically improve a RAG app. Often it just gives the model more irrelevant context.
• To fix bad RAG answers, measure retrieval quality, rerank the results and send only the most relevant context to the model.
• Giving an agent every tool (for example 30 of them) makes it less reliable, not more capable.
• Give each agent only the tools it needs and control exactly when those tools can be called.
• Using the biggest model for everything hurts margins and latency.
• Send simple tasks to smaller models and save the expensive large models for tasks that really need them.

Resources mentioned

• BASWE.Ai Engineer (Skool community) · community · skool.com · paid
  The creator's paid community and program, with an AI learning roadmap (including the full ops and evaluation track), daily calls with engineers and recruiters, resume and portfolio help, and a job-search pipeline.
  Also in: Basic RAG Pipeline in 60 Seconds: From Documents to Grounded Answers (Bashiri Smith on Facebook · notes), Pointer to Bashiri Smith's Complete AI Engineer Roadmap for 2026 (Bashiri Smith on Facebook · notes), Step-by-Step Roadmap to a $200K+ AI Engineering Role (Bashiri Smith on Facebook · notes), How to Evaluate a RAG Pipeline: Retrieval vs. Generation (Interview Answer) (Bashiri Smith on Facebook · notes) and 76 more

Try this

• Measure retrieval quality in your RAG app before changing how many chunks you retrieve.
• Add a reranking step and send only the most relevant chunks to the model.
• Cut each agent's toolset down to what it needs and control when each tool can be called.
• Set up model routing: smaller models for simple tasks, large models only when needed.
• Comment "JOIN" to get the community link (promotional call to action).