ASCI Logo
ASCI
Programs
Practice DSA
Degrees & Careers
Pricing
Log inGet Started

Quick Navigation & Search

Search curriculum tracks, DSA problems, systems theory and tools...

agentic ai

Curriculum Index

Current LevelInitiate
0/ 500 XP
Gravit Certificate:
100% Required
  • 1.1 The Shift from Prompting to Agency
  • 1.2 The Three-Layer Agent Architecture
•
1
2
3
4
5
6
7
8
9
10
Module 1: Foundations of Agentic AI & The PRAL Loop•Lesson 2

1.2 The Three-Layer Agent Architecture

### Decoupling the Agent Stack
Every production agent relies on three decoupled subsystems:
* **Perception Layer:** Context summarization, multimodal vision/audio transcription, and state representation.
* **Cognition Layer:** Planning heuristics, memory lookup (vector + episodic key-value store), and decision gating.
* **Action Layer:** Structured JSON tool schema calling, deterministic API execution, and safety validations.

example.pyPython 3.11
class Perception:
    def parse_input(self, text: str) -> dict:
        return {"intent": "SEARCH", "query": text}

p = Perception()
print(p.parse_input("quantum computing")["intent"])
Python 3.11
•
Interactive Challenge
+75 XP
Challenge Objective

Implement parse_intent to inspect queries and return 'SEARCH'.

solution.py
UTF-84 SpacesAuto-saved
cpython3.11:~/workspace
CPython 3.11.8 (WASM)
$python3 -u solution.py
Click "Compile & Run" (or press Ctrl + Enter) to compile AST, verify bytecode, and execute the solution.