AI Service Design — Part 2: When to Use ML vs LLM
When working on AI-driven products, one of the most common decisions is whether to use traditional ML or LLM-based approaches.
LLMs have advanced to the point where they can solve almost anything—but that doesn’t mean they’re always the right choice.
In this video, we break down how to decide between ML and LLM depending on the nature of the problem.
Topics covered:
Strengths of using ML
How to decide when ML is the better fit
How to decide when LLMs make more sense
Framing problems as deterministic vs interpretive
Timestamps
00:00 Intro
00:03 Advantages of using ML
03:14 Deciding when to use ML
04:05 Deciding when to use LLM
05:42 Deterministic vs interpretive problems