Data Engineer

Maxime Beauchemin Portfolio

Founder at Preset

GitHub profile of Maxime Beauchemin, original creator of Apache Airflow and Apache Superset — two of the most widely deployed open-source tools in data engineering — and founder of Preset, whose 2026 activity centers on Agor, a Preset-hosted workspace for running coding agents on isolated git branches (1.4k stars).

Apache AirflowApache SupersetData OrchestrationOpen Source

Location

San Mateo, CA

Companies

Preset

Skills

Apache AirflowApache SupersetData OrchestrationOpen Source

Notable

  • Creator of Apache Airflow (47k GitHub stars) and Apache Superset (74.9k GitHub stars)
  • 8,124 GitHub contributions in the last year
  • Agor (preset-io): 1.4k GitHub stars, 2,663 commits

What makes it work

A breakdown of the choices that make this portfolio stand out.

1

Two category-defining tools originating from one profile

Airflow (orchestration) and Superset (visualization) each anchor a different layer of the modern data stack, and both trace back to the same person's early work — a rare breadth of foundational open-source contribution.

2

Current work in the open, not just legacy credits

The contribution feed is dominated by Agor, a Preset project for running Claude Code, Codex and Gemini agents on isolated git branches: 103 commits and 91 pull requests in September 2026 alone. The pinned awesome-data-engineering repo, by contrast, is a fork of someone else's list with no commits since 2017 — the live work is the better signal.

3

New, smaller experiments alongside the legacy flagship projects

claudette-cli, a git-worktree manager that spins up isolated Superset dev environments ready for Claude Code, shows ongoing hands-on building rather than resting on Airflow and Superset's reputation. duckstreams (Kafka topics as DuckDB virtual tables) is so far a README-only concept with three commits.

What Data Engineers can take from this

Specific, actionable tips to apply to your own portfolio — no generic advice.

  • If you've built a foundational tool, keep a visible, updated presence (new small repos, active involvement) rather than letting the profile read as a historical artifact.

  • Audit your pinned repos: a dormant 2017 fork at the top of a profile undersells the work happening in the contribution feed below it.

  • Let your profile show tools spanning different layers of the stack you work in, not just one — it demonstrates systems-level thinking, not a single point solution.

Projects

Maxime Beauchemin's bio line carries two of data engineering's defining tools, but the commit graph is somewhere else entirely: 103 commits to Agor, an agent workspace, in September 2026 alone. The legacy sits in the header and the current work sits in the feed.

Open SourceadvancedOngoing since 2015

Apache Superset

PythonReactSQLApache Software Foundation

An open-source data exploration and visualization platform, first published under the name Caravel (his original fork still carries that description: "a data exploration platform designed to be visual, intuitive, and interactive"). Now an Apache top-level project and the basis for Preset.

Taking a project from a personal repo to an ASF project with a commercial company built on it covers the whole arc from code to community to business.

  • Bio lists him as creator

  • claudette-cli exists specifically to speed up Superset development

Open SourceadvancedOngoing

Apache Airflow

PythonDAGsWorkflow orchestration

"A platform to programmatically author, schedule, and monitor workflows" — the orchestrator whose DAG-as-Python-code model became the default mental model for batch data pipelines. Now an Apache Software Foundation project.

Airflow's core abstractions (DAGs, operators, schedules, backfills) are the vocabulary most data engineering interviews still use.

  • Bio lists him as creator

  • One of two category-defining tools from the same profile

SaaS ProductadvancedOngoing (2026)

Agor

TypeScriptpnpm/TurborepoDocker ComposePostgreSQLMCP

A self-hosted, multiplayer-ready web workspace for running coding agents (Claude Code, Codex, Gemini and others) on isolated git branches. Each branch is a working directory with its own dev environment and conversation history, and the workspace adds per-prompt token and cost accounting and an MCP endpoint the agents can drive themselves.

It shows a data-tools founder applying the same orchestration instincts (isolation, observability, cost tracking) to agent workflows.

  • Branches shown as cards on a board

  • Brings no model of its own; drives the agent CLIs and SDKs users already have

  • Optional Unix-level isolation for teams

Ready to build your portfolio?

Follow the Data Engineer roadmap — skills, projects, and timeline to get hired.

Data Engineer Roadmap