Agentic Data Science Workshop
Compress 2 weeks of work into 1 day with coding agents. A fundraiser for SciPy 2026.
Your donation funds SciPy's Financial Aid program, which brings 15–20 students from around the world to the conference—flights, lodging, and full attendance—who wouldn't otherwise be able to attend. For many, it's a transformative, career-launching experience.
Register your interest; we'll confirm by email and reserve your spot. Then donate and upload your receipt within 48 hours.
How it works
- 1Register interest
- 2We confirm and reserve your spot
- 3Donate $1,000+ or $200 (students) and upload receipt within 48h
- 4We verify and enroll you
A quick word from Eric
What you'll learn—and where your donation goes. You'll learn to compress two weeks of work into one day using coding agents, with battle-tested practices you can use the next day. Your donation goes straight to NumFOCUS for SciPy financial aid; it's tax-deductible and no money passes through Eric's hands.
Level up your productivity and skills—and help us reach our $35,000 goal for SciPy financial aid so more people can join the Scientific Python community.
Why This Workshop?
5-10x productivity gains
Compress two weeks of analysis work into half a day by leveraging coding agents for implementation while you focus on inquiry.
Battle-Tested Techniques
These techniques have been proven and are being refined in real-world data science practice.
Immediate application
Walk away with skills and agent configurations you can use in your work the very next day.
What You'll Learn
Prescriptive prompting for experiments
Frame questions as observable outcomes that agents can verify.
Project structure for agents
Create predictable file layouts so agents know where to put things.
Logging & verification strategies
Build introspectable experiments that agents can debug autonomously.
Report writing with AI
Generate human-readable summaries with explicit scrutiny of outputs.
Staying in control during EDA
Gate analysis on plots, structure sessions with journals, keep the human in the driver's seat.
The minimalist version first
Prove pipelines work with tiny models before scaling up.
Your Instructor
Eric J. Ma
Eric is a Data Scientist at Moderna, where he leads two data science teams in the preclinical and clinical space. He's been using coding agents to compress experimental timelines from days to minutes.
He's the author of several open-source packages including pyjanitor and nxviz, a contributor to PyMC, and has taught tutorials at SciPy, PyCon, and ODSC.
His blog posts on agentic data science have helped hundreds of data scientists adopt these techniques in their own work.
Frequently Asked Questions
Is the donation tax deductible?▼
Yes! All donations go directly to NumFOCUS, a 501(c)(3) nonprofit. You'll receive a receipt for tax purposes.
What if I can't attend live?▼
Video recordings will be available to all confirmed participants. You won't miss anything.
Do I need prior AI experience?▼
Basic Python and data science experience is assumed. No prior AI/LLM experience required—we'll cover everything you need.
What coding agents will we use?▼
The techniques work with Cursor, Claude Code, OpenCode, and similar tools. You'll learn patterns that transfer across platforms.
When are the sessions?▼
March 27 and April 3, 2026, from 2:00-3:50 PM EST. Two sessions, two hours each. Video recordings will be available to confirmed participants.
How do I confirm my donation?▼
We'll email you and reserve your spot. Then you donate at least $1,000 (or $200 if you're a verified student) and upload your receipt within 48 hours. We verify and enroll you, and you'll get a confirmation email.
Is there a student rate?▼
Yes — $200 for verified students. When you register, you can request the student rate. We prefer you use a school email so we can verify quickly without storing ID photos. If you don't have a school email, you can upload a photo of your student ID (name and school visible) and we'll verify it.
If your question isn't answered here, send me a message.