Koudai Aono koxudaxi

Python · Developer Tools · OSS

Koudai Aono (青野高大)

I build and maintain Python tools that turn schemas into typed code and make new language features usable in real projects.

Creator and lead maintainer of datamodel-code-generator. Co-author of PEP 750, which brought template strings to Python 3.14. Author of Python IDE tooling with 1.2M+ downloads.

Open to a full-time role, bringing datamodel-code-generator in-house with me, or collaboration, especially in developer tools and production-quality OSS.

  • 14M+ monthly PyPI downloads
    datamodel-code-generator
  • 6.5k GitHub stars
    across my OSS projects
  • 1.2M+ downloads
    my JetBrains plugins
  • Since 2019 creator and lead maintainer
    datamodel-code-generator

Based in Tokyo, Japan. Available for global remote work.

Get in Touch See Public Impact Sponsor

Independent Adoption

Public projects use datamodel-code-generator in SDK and model generation, developer tooling, configuration workflows, and test or evaluation infrastructure. Each project below links to the code or documentation for that use.

AI and Agent Systems

OpenAI Codex SDK MCP Python SDK Modular MAX E2B Cognee

Data and Platform Infrastructure

Apache Airflow Airbyte OpenMetadata

Observability and Product Infrastructure

Datadog OpenTelemetry PostHog

Tests and Evaluation

vLLM browser-use DSPy

See all public dependents

How I Build OSS

I usually start with a repeated problem that existing tools do not solve well. I define the model, build the tool, add the integrations and documentation needed to use it, and keep it compatible as Python and its libraries change.

Schema to Typed Python

I created datamodel-code-generator in 2019 and have been its lead maintainer since. It now lives in its own GitHub organization. It turns OpenAPI, JSON Schema, GraphQL, CSV, and other inputs into Pydantic v2 models, dataclasses, TypedDicts, and msgspec structs. Since 0.84, it also generates FastAPI server and HTTP client packages.

The work covers schema semantics, naming, compatibility, generated-code quality, CI usage, and migration across Python and model-library versions. It is also documented for AI coding agents.

Repository · Documentation · llms.txt · coding-agent skill · fastapi-code-generator

Python Language and Typing

I co-authored PEP 750, which introduced template strings in Python 3.14. Their interpolations stay structured so tools can inspect and transform them before rendering. My talks examine how typing concepts hold up in maintained Python projects, including decorators and the difference between missing, None, and unset values.

My current work explores linting, safe rendering, and structured-data tools built on t-strings.

PEP 750 · t-strings.help · t-linter · tstring-html

Adoption in the IDE

I built JetBrains plugins when widely used Python tools still lacked first-class IDE support. The Pydantic, Ruff, and Poetry plugins have 1.2M+ combined downloads.

The Poetry plugin was later bundled with PyCharm. The Ruff plugin preceded JetBrains' native Ruff support.

Pydantic plugin · Ruff plugin · Poetry plugin

If your team relies on any of these projects, sponsoring is the most direct way to support long-term maintenance, compatibility updates, releases, and documentation. Sponsor on GitHub

Where This Fits

This work is most useful to teams building AI coding systems, SDK and API platforms, schema infrastructure, or Python developer tooling. I can own a problem area, choose the technical direction, ship the work, and maintain the public project over time.

What I Am Looking For

I am exploring new opportunities, especially in developer tools and production-quality OSS.

That can be a full-time role, bringing datamodel-code-generator in-house with me, or collaboration.

The company sets product goals, priorities, and constraints. I take ownership of problem discovery, technical direction, delivery, adoption, and long-term maintenance within that context.

Based in Tokyo. I previously worked fully remote with a US company and at the Tokyo branch of a UK company. Employment history is available on LinkedIn.

Get in Touch LinkedIn GitHub Sponsor

Bringing datamodel-code-generator in-house

If your company builds on datamodel-code-generator, I am open to discussing bringing the project in-house with me as its lead maintainer.

Selected Talks

My talks focus on bringing typing and language concepts into maintained, real-world Python projects.