Workflows vs agents
Comes up in: AI engineering, AI system design
This concept is being written. It gets the full five steps — explanation, predict, interview angle, practice with tests, and review cards. Here is exactly what it will cover, so you can study it elsewhere in the meantime.
What it covers
- definitions (workflow: code-orchestrated LLM calls
- agent: model-directed loop with tools)
- patterns: prompt chaining, routing, parallelisation, orchestrator-workers, evaluator-optimiser (cite Anthropic "Building effective agents")
- ReAct (Yao et al. 2022)
- start simple
- cost/latency/error trade-offs
Words you will know
What you will practise
- solve one task as a chain, a router and an agent
- compare evals and cost
Interview questions this answers
- What is the difference between a workflow and an agent? When would you build an agent?
- Describe the evaluator-optimiser pattern.
- What is ReAct?