We’re sunsetting JetBrains DataSpell and will no longer be developing it as a standalone IDE. For data science-related workflows, we recommend PyCharm.
Learn more about how this affects your license and access to the product here: jb.gg/eq6wah
Thanks so much to everyone who used, supported, and believed in DataSpell!
Every data team exploring agentic analytics hits the same question: Should we build this ourselves or use a platform?
At first, building looks simple. You connect an LLM to your data, add some docs, and it works.
But the hard part is making it reliable over time, across teams, and in production.
That’s where DIY breaks: unclear business logic, inconsistent answers, growing maintenance.
What looked simple becomes a system to maintain.
There are several ways to connect an AI agent to your data. At first glance, all three approaches give you the same thing : answers in natural language.
But the real question is: Can you trust those answers?
It usually comes down to this:
Direct access → fast, but requires constant verification
Semantic layer → reliable, but slow to build
Automated layer → combines speed with built-in review for reliability
The data analyst role is changing. In the dashboard era, analysts wrote queries and built charts.
In the agent era, the job becomes defining metrics, building semantic contracts and designing guardrails for AI systems.
🧵
In Part 2 of “can you trust your data agent?”, we explore the emerging AI analytics stack, metrics-as-code and Git governance. All the elements at the foundation behind Databao at JetBrains.
Read more: jb.gg/f0xh8l
Two analysts can answer the same question and get two different answers. Now imagine AI agents doing the same thing.
That’s the agent swarm problem in a nutshell. Agentic analytics doesn’t need smarter agents – it needs a shared semantic foundation.
At JetBrains, this is exactly the problem we’re trying to solve with Databao.
We’re working on a context engine and data agents to build trustworthy, replicable agentic analytics.
Read more here: jb.gg/emrvqt
When building AI agents, people tend to run into the same issues:
🫠 It kind of works, but I don’t know why.
🫠 The agent is hard to extend.
🫠 It’s hard to debug.
🫠 It’s hard to trust.
Check a free step-by-step course by @t_redactyl on building AI agents with Python, #LangGraph, #MCP, #Ollama:
youtu.be/j4sNAwrx3kc?si=ae-w…
Want to move beyond clever prompting? 👀
This tutorial shows how to design and install agent skills – modular task workflows that reduce context clutter and improve consistency.
From structured `SKILL.md` files to installing best-practice Python Skills, it’s all covered.
▶️ Watch the video: jb.gg/agent-skills
Guja is an analytics engineer at Carnival Maritime. We spoke with him about what drew him to Databao and how it helped speed up his ad-hoc analytics work in a complex data environment.
Read his story: jb.gg/2868g8
A full day of high-quality #Python content – free to watch, anytime.
6 hours of talks on:
• Modern software development
• Data and AI workflows
• Real-world engineering practices
Whether you use Python daily or are just getting started, you’ll find sessions covering all parts of the stack.
Watch now: youtube.com/watch?v=qKkyBhXI…
The free online Python conference hosted by @pycharm is live on YouTube!
Join "Python Unplugged on PyTV" with Carol Willing, Paul Everitt, Sheena O’Connell, and other people behind the tools and libraries you use every day.
1 day · 6 hours of live talks · 15 speakers
Hop in: youtube.com/watch?v=qKkyBhXI…
Looking to integrate DataGrip into your workflow that already includes other JetBrains IDEs?
Vladimir Kuznichenkov, a Staff Infrastructure Engineer at QDO, shares how his team integrated DataGrip with GoLand for schema validation, testing, and data discovery.
Discover the Database and SQL plugin in any JetBrains IDE: jb.gg/postgreVKx
Discover how Omar Sarhan, a Data Engineer at Geolytix, helps developers and data scientists work with geospatial data to create models. With DataGrip, he has the toolkit he needs to manage code and view the entire database.
Start a free 30-day trial or use DataGrip for non-commercial purposes for free forever: jb.gg/postgreOMx
Hear from Lukasz Galezewski, a Senior Software Engineer at Allegro, on how he uses DataGrip to work with MySQL, PostgreSQL, and Oracle databases at Poland’s biggest marketplace.
Discover DataGrip, your powerful cross-platform IDE for relational and NoSQL databases: jb.gg/postgreLGx
A new milestone for our data products family!
Databao Agent ranked #1 in the Spider 2.0–DBT benchmark. We achieved this by treating the agent like a colleague: providing clearer context and implementing a reliable workflow.
Learn more → jb.gg/n1jruu