Data science.
On your terms.

From a question to an experiment you can trust. Aner brings data, models, and reproducibility into one programming language designed for data science.

Explore the language Be part of what’s next

Native Aner runtimeOpen source release ahead

Native Aner

import aner.dataset
import aner.nn
import aner.ml

let split = dataset.iris().split(test: 0.2, seed: 2026)
let network = nn.sequential(seed: 42)
    .dense(inputs: 4, outputs: 8)
    .tanh()
    .dense(inputs: 8, outputs: 3)

let model = ml.classifier(network, standardize: true)
let fitted = model.fit(split.train(), epochs: 2000, rate: 0.1)
print(fitted.evaluate(split.test()).summary())

Scaling fitted on training data. A separate test set.

SMALL MODEL. CLEAR STRUCTURE.4 inputs → 8 hidden → 3 classes

Illustrative topology · Run the example in Aner

BUILT AROUND YOUR WORK Data with context Reproducible experiments Models you can inspect

A MORE CONNECTED WORKFLOW

Less glue.
More science.

Data science should make the question clearer. Aner is being designed to reduce the steps between understanding your data and understanding your result.

01 / UNDERSTAND

Know what goes in.

Explicit schemas, visible missingness, and bounded CSV processing. Make assumptions part of your workflow.

Work with data

02 / EXPERIMENT

Build a clear experiment.

Compose a neural network, fit preprocessing on training rows, and evaluate a separate test dataset.

Explore real examples

03 / UNDERSTAND AGAIN

Look beyond the score.

Inspect recorded training graphs and results. Carry variables between notebook cells in VS Code.

See visual debugging

ONE LANGUAGE. ROOM TO GROW.

A toolkit that
speaks data science.

7 modules share native types and a coherent workflow. Start with today’s CPU implementation; help shape what comes next.

Meet the modules

aner.nn

Describe the model. See the learning.

aner.data

Know what enters your analysis.

aner.dataset

A dataset with its context attached.

aner.ml

A shared workflow for learning.

RESEARCH, BEFORE SHORTCUTS

Better tools begin
with better questions.

Aner started with the friction researchers face when working with complex datasets. We are exploring how a language can make assumptions, resource use, and experimental state easier to see.

Research software in active development. Working features and proposed designs are documented separately.

Explore the research direction

HELP SHAPE ANER

Good science is collaborative.
So is a good language.

Bring your dataset challenges, teaching experience, or engineering skills. There’s room for your perspective here.

Join the conversation

Anerita, Aner’s friendly community mascot

Meet Anerita. Your curious companion.

Data science.On your terms.

Less glue.More science.