# WecoAI/aideml

AIDE: an LLM agent for machine learning engineering - the research Weco grew out of. Referenced in OpenAI MLE-bench.

Repository: https://github.com/WecoAI/aideml
Canonical: https://ross.abutalabs.com/products/aideml
Homepage: https://weco.ai
Language: Python
License: MIT
License Family: permissive
Topics: ai, data-science, llm, machine-learning, ai-agents, automated-machine-learning, autonomous-agents, autoresearch, code-optimization, self-improving-ai
Last push: 2026-08-17T11:23:43+00:00

## Health v2 (maintenance only)
Score: 65/100 (v2, computed 2026-09-02T17:46:02.011165+00:00)
- activity 98, release rhythm 23, longevity 63
- inputs: {"age_days": 882, "days_push": 16, "days_rel": 301, "gap_med": 286, "n_releases_24m": 2}
- flags: none
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1494, forks 221 (observed 2026-08-28T04:04:52.771096+00:00)

## What it is
AIDE ML is an open-source Python package implementing an LLM-driven agent that uses agentic tree search to autonomously write, debug, and improve machine-learning code against a user-defined metric. It is the research reference build behind Weco and was referenced in OpenAI's MLE-bench.

## Use cases
- automatically build ML pipelines from a dataset and a natural-language goal
- enter Kaggle-style competitions with an autonomous AI data scientist
- research new agent architectures by swapping search heuristics or LLM backends
- optimize code against a measurable metric using tree search
- replicate the AIDE paper's results on MLE-bench
- prototype predictive models without hand-tuning

## When to choose
- you want an autonomous agent to draft and iteratively improve ML code for a dataset
- you're an agent-architecture researcher wanting an extensible tree-search baseline
- you need a research-friendly, pip-installable implementation of the AIDE algorithm

## When to avoid
- you need a production ML platform with experiment tracking and dashboards - use the Weco product instead
- your task has no quantifiable evaluation metric
- you want a no-code AutoML tool with a GUI

## Facets
- artifact type: library
- maturity: active
- function: agent-framework, machine-learning, llm-inference, cli, data-science
- domain: machine-learning, artificial-intelligence, data-science
- platform: python, cli, cross-platform
- tags: llm-agent, tree-search, automated-machine-learning, kaggle, autoresearch, code-optimization, mle-bench, ai-agents, automation

## Member repositories
- WecoAI/aideml (main) score 65

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:04:52.771096+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-30T04:33:20.642325+00:00, confidence not recorded.
  - readme: https://github.com/WecoAI/aideml (fetched 2026-08-28T04:04:52.771096+00:00, sha ec06ce25630a)
  - homepage: https://weco.ai (fetched 2026-08-29T11:38:49.190351+00:00, sha 764e447407cb)
  - site_page: https://docs.weco.ai/ (fetched 2026-08-29T11:38:49.192909+00:00, sha ce9ba76906ff)
  - site_page: https://docs.weco.ai (fetched 2026-08-29T11:38:49.195018+00:00, sha ce9ba76906ff)
  - registry_pypi: https://pypi.org/pypi/aideml/json (fetched 2026-08-29T11:38:49.197345+00:00, sha d54b8110ed13)
- Data as of 2026-08-30T08:39:29.467469+00:00.