About this event
Can an AI agent actually make you a better data scientist, or just a faster one with messier models?
Agents can now write your pipelines, generate your features, even tune your models. The real question is what happens to quality when they do. Who's checking the work?
That's what this session is about. No slides full of theory. Marie takes one real data science use case and works through it live, so you can see exactly:
- Where an agent genuinely helps - and the steps where you're still better off doing it yourself
- What it generates at each stage - from data prep to model evaluation, shown on screen, not described
- The guardrails that keep your models robust and reliable - so you ship faster without shipping mistakes
Who's speaking:
Marie Sacksick spent 10 years in data science, first building models, then managing the teams that build them. Two years ago she joined Probabl, the scikit-learn company, to help develop the foundational tools data scientists around the world use every day. She's seen both sides: what agents promise, and what they break.
Who this is for:
Data scientists and DS team leads who want to use agents in real work, not demos, and stay in control of what gets shipped.
It's live, and the Q&A is the best part. Bring your questions.
📅 August 27 · 12pm EST / 6pm CEST · 45 minutes
Probabl
The Tabular AI company by the creators of scikit-learnProbabl is the Tabular AI company by the creators of scikit-learn. After standardizing how the world does machine learning, we're enabling enterprise data science teams to adopt AI with confidence.
