CrowdCent Challenge

Open data science competitions for ML engineers and data scientists

The CrowdCent Challenge is an open data science competition designed for machine learning engineers, data scientists, AI agents, and other technical professionals to hone their skills in a real-world setting.

What is CrowdCent?

CrowdCent is on a mission to decentralize investment management by changing the way investment funds make decisions and allocate capital. We are the machine learning and coordination layer for online investment communities looking to turn their data into actionable, investable portfolios.

Installation

Using uv (Recommended)

uv add crowdcent-challenge

Using pip

pip install crowdcent-challenge

Quick Start

Get an API Key: Generate your key from your profile settings

Set up authentication:

export CROWDCENT_API_KEY=your_api_key_here
# or create a .env file with: CROWDCENT_API_KEY=your_api_key_here

Start competing:

from crowdcent_challenge import ChallengeClient

# Initialize client for a challenge
client = ChallengeClient(challenge_slug="hyperliquid-ranking")

# Download training data
client.download_training_dataset("latest", "training_data.parquet")

# Download inference data
client.download_inference_data("current", "inference_data.parquet")

# Submit predictions, from a parquet or a CSV file
client.submit_predictions(file_path="predictions.parquet")

# Every file comes and goes as CSV too: just name the path .csv
client.download_inference_data("current", "inference_data.csv")
client.submit_predictions(file_path="predictions.csv")

# Or test a new model with an experimental submission + private note
# (requires another slot to have a non-experimental submission)
client.submit_predictions(
    file_path="predictions.parquet",
    slot=2,
    is_experimental=True,
    notes="2-layer transformer w/ sector embeddings",
)

Available Challenges

Hyperliquid Ranking: Rank crypto assets on Hyperliquid by expected relative returns
Equity NLP: Coming soon!

CLI Usage

The package includes a command-line interface:

# List all challenges
crowdcent list-challenges

# Set default challenge
crowdcent set-default-challenge hyperliquid-ranking

# Download data
crowdcent download-training-data latest -o training.parquet
crowdcent download-inference-data current -o inference.parquet

# Submit predictions
crowdcent submit predictions.parquet

# Submit an experimental prediction with a private note
crowdcent submit predictions.parquet --slot 2 --experimental \
    --notes "2-layer transformer w/ sector embeddings"

# Scores, and what your API key may do
crowdcent performance
crowdcent whoami

# Backtest the meta-model
crowdcent sim run --config '{"n_long": 10, "n_short": 10, "rebalance_days": "10t"}'

# Live trading: preview, then execute the plan (testnet unless --network mainnet)
crowdcent trade preview
crowdcent trade execute PLAN_HASH

# CrowdCent Cloud: create, run, and schedule a project
crowdcent cloud create "My model" --recipe hyperliquid-ranking --challenge-access
crowdcent cloud run PROJECT_ID --wait
crowdcent cloud schedule PROJECT_ID --trigger on_inference_release --release-challenge hyperliquid-ranking

Every command prints JSON; run crowdcent <command> --help for its options.

Documentation: docs.crowdcent.com

AI Agents / MCP Server

This package ships with a built-in Model Context Protocol (MCP) server, so assistants like Claude and Cursor can download data, submit predictions, backtest strategies on the meta-model, and check your performance in natural language. Connect the hosted server at https://mcp.crowdcent.com/mcp with your API key, or run it locally with one line:

{"mcpServers": {"crowdcent": {
  "command": "uvx",
  "args": ["--from", "crowdcent-challenge[mcp]", "crowdcent-mcp"],
  "env": {"CROWDCENT_API_KEY": "your_api_key_here"}
}}}

See the AI Agents guide for hosted setup, the simulation tools, and live trading.

Contributing

Contributions are welcome! The crowdcent-challenge client library and documentation are open source.

See our contributing guidelines for details on:

Forking and cloning the repository
Setting up development environment
Making changes and submitting PRs

Have Questions?

Release files for crowdcent-challenge 0.2.14

For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.

Source distribution (sdist)

Source distribution for crowdcent-challenge 0.2.14
File	Size	Uploaded
crowdcent_challenge-0.2.14.tar.gz	504.5 kB	Sep 25, 2026	Details

Built distribution (wheel)

Table of built distributions (wheels) for crowdcent-challenge 0.2.14
File	Interpreter	ABI	Platform	Reset
crowdcent_challenge-0.2.14-py3-none-any.whl 70.4 kB Sep 25, 2026	Python 3	none	any	Details

Total release size: 574.9 kB

Release files / crowdcent_challenge-0.2.14.tar.gz

Download URL	crowdcent_challenge-0.2.14.tar.gz
Size	504.5 kB
Tags	Source
SHA-256 checksum How to use checksums	`ca1b50444d8c1f85839f9c86b7b2ad1f95e9ae45b9c33ddf3911864676eeeed1`
BLAKE2b-256 checksum How to use checksums	`436d02b9b0c886f7d90207a69c636d87b28ef510ea9b748bd6650fd62a63acc0`
Upload date	Sep 25, 2026
Uploaded using Trusted Publishing? What is trusted publishing?	No
Uploaded via	`uv/0.7.13`

Release files / crowdcent_challenge-0.2.14-py3-none-any.whl

Download URL	crowdcent_challenge-0.2.14-py3-none-any.whl
Size	70.4 kB
Tags	Python 3
SHA-256 checksum How to use checksums	`43984d43599f855b1ef83bbf1c96942ddeaa0fb2dca973e6dfa5a29b3207e629`
BLAKE2b-256 checksum How to use checksums	`8502d737c77fdc7ff2a85219df7c6c8255c08e15f20365f251d3585979780242`
Upload date	Sep 25, 2026
Uploaded using Trusted Publishing? What is trusted publishing?	No
Uploaded via	`uv/0.7.13`