mlane/llm-engineering-cheatsheet

Timeless principles and best practices for working with language models - tooling-agnostic, future-proof, and clear.

27
/ 100
Experimental

Covers universal prompting patterns (zero-shot, few-shot, role-based, constrained output) and the five-component prompt anatomy (role, task, input, constraints, examples) applicable across any model provider. Emphasizes treating LLMs as probabilistic next-token predictors rather than deterministic systems, with practical guidance on temperature/sampling controls, token budgeting, and systematic debugging via failure mode analysis. Includes a minimal Python example using OpenAI's API but remains framework-agnostic, grounded in principles that predate current tooling.

No Package No Dependents
Maintenance 6 / 25
Adoption 5 / 25
Maturity 9 / 25
Community 7 / 25

How are scores calculated?

Stars

10

Forks

1

Language

License

MIT

Category

optimization

Last pushed

Dec 24, 2025

Commits (30d)

0

Get this data via API

curl "https://pt-edge.onrender.com/api/v1/quality/prompt-engineering/mlane/llm-engineering-cheatsheet"

Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.