| Below is from Aurimas Griciลซnas showing hierarchical layers for AI Agents Sometimes ๐๐ ๐๐ด๐ฒ๐ป๐๐ might seem simple from the outside. They are not ๐
There are many moving pieces that production ready Agentic Systems rely on.
I like to split the ๐๐ถ๐ฒ๐ฟ๐ฎ๐ฟ๐ฐ๐ต๐ ๐ผ๐ณ ๐ก๐ฒ๐ฒ๐ฑ๐ of an AI Agent according to the maturity of the product:
๐๐๐พ.
We usually start building and experimenting with Raw Model APIs. These already rely on complex underlying infrastructure.
โก๏ธ ๐๐๐/๐๐๐ Resources.
โก๏ธ ๐๐ข๐ด๐ฆ ๐๐ฏ๐ง๐ณ๐ข๐ด๐ต๐ณ๐ถ๐ค๐ต๐ถ๐ณ๐ฆ that orchestrates Model deployment. Think Kubernetes, Slurm, vLLM.
โก๏ธ ๐๐ฐ๐ถ๐ฏ๐ฅ๐ข๐ต๐ช๐ฐ๐ฏ ๐๐ฐ๐ฅ๐ฆ๐ญ๐ด themselves that required tens of millions of dollars to be trained.
๐๐๐.
To produce a reliable MVP you will need some internal data and stability to your system. This will require you to make choices in how you:
โก๏ธ ๐๐ต๐ฐ๐ณ๐ฆ the data: Vector DBs, Graph DBs etc.
โก๏ธ ๐๐ณ๐ค๐ฉ๐ฆ๐ด๐ต๐ณ๐ข๐ต๐ฆ the system: LLM Orchestration frameworks help with solving issues like retries, chaining your prompts, tool calling etc.
๐ฝ๐๐ฉ๐.
Scaling up for Beta requires more stability and insights into what is happening inside of the system.
โก๏ธ ๐๐ฐ๐ฅ๐ฆ๐ญ ๐๐ฐ๐ถ๐ต๐ช๐ฏ๐จ helps in choosing best LLMs for your prompts, prompt management, fallback mechanisms in case of unresponsive LLM APIs or hitting API limits. Etc.
โก๏ธ ๐๐๐ ๐๐ฃ๐ด๐ฆ๐ณ๐ท๐ข๐ฃ๐ช๐ญ๐ช๐ต๐บ allows you to see into the actions your system is performing. You will need this for debugging and evolving your Agents.
๐๐ผ.
To expose your application to the general public you would need additional automations and guardrails to prevent disasters that could shut your business in seconds.
โก๏ธ ๐๐ท๐ข๐ญ๐ถ๐ข๐ต๐ช๐ฐ๐ฏ helps you to go beyond passively observing the system to proactively monitoring it by bringing automation via evaluation rules that are applied against steps performed by your Agents.
โก๏ธ ๐๐ฆ๐ค๐ถ๐ณ๐ช๐ต๐บ is critical to avoid disasters related to data leakage. Think Guardrails, Red Teaming etc.
โ๏ธAnd these are just basic requirements, there is more ๐
Come in the emerging requirements:
โก๏ธ Memory.
โก๏ธ Computer Use.
โก๏ธ Complex integrations. (MCP!)
โก๏ธ ...
| 
