Work through large tasks one step at a time You let the actor run one observable step, inspect compact evidence, and continue from live state. This avoids stuffing a large task into one prompt or generated script. python academy academy/topics/rlm-pipeline website/content-src/academy/course.mjs academy Work through large tasks one step at a time
ax
Python axllm
TypeScript Python Java C++ Go Rust
Unit 6 · Solve long and complex tasks

Work through large tasks one step at a time

You let the actor run one observable step, inspect compact evidence, and continue from live state. This avoids stuffing a large task into one prompt or generated script.

agent()9 focused minutesNot started
Unit example (nearest native match)

See the idea in context

investigator = agent(signature, {
    "contextFields": ["logs"],
    "contextPolicy": {"preset": "lean", "budget": "balanced"},
    "runtime": {"language": "JavaScript"},
})
  1. Filter inside the runtime

    The full records stay available to code instead of being repeated in a prompt.

  2. Expose compact evidence

    Logging only the count gives the next turn a useful observation.

  3. Continue from live values

    Later turns can reuse matches without recomputing or replaying the dataset.

Run itIn your own project
pip install axllm

from axllm import ai, ax

llm = ai('openai', api_key=os.environ['OPENAI_API_KEY'])
classify = ax('review:string -> sentiment:class "positive, negative, neutral"')

Set OPENAI_APIKEY in your environment before running provider-backed code.

In the ax repo

From a clone of the ax repo:

npm run example -- python src/examples/python/long-agents/incident-log-forensics.py
Active practice

Show that you can use it

Answer 2 in a row to learn this · attempt 1
Keep exploring

Source-backed follow-up