Data Berlin #32
Fresh season, fresh perspective and a brand new website.
🐣 Easter just passed; a small pause that often gives a bit of space to reset.
Some time ago we asked what we could do to improve this newsletter. The answer was more job posting. Unfortunately, Substack space is limited, so we had to come up with a better idea.
We built a dedicated job board for real data and AI roles in Berlin. You can find the roles in this issue and more on the databerlin.net page.
If you’re hiring or know a company we should add, just reach out to us.
If you’ve been thinking about a change - or even just staying aware of the market - this is a good time to take a look around.
📖Interesting Readings
With dlt’s AI workbench, the author automated schema discovery for an undocumented finance platform by sampling ODBC tables into DuckDB, inferring types and naming semantics, building a business ontology, and validating join paths in hours instead of weeks. – Reverse-Engineering a Black-Box ERP with Claude Code
WITH clauses in PostgreSQL are no longer automatic performance fences: since v12 they may be merged or cached based on usage, volatility, recursion, and side effects, so choosing between inline execution, explicit caching, or temp tables is a deliberate performance decision. – Good CTE, bad CTE
Even with agentic tooling, understanding core loading patterns remains table stakes: picking rebuild-all vs change-only updates shapes cost, reliability, and scalability long before any automation layer helps. – Full Refresh vs Incremental Pipelines
An interview-style thought experiment argues that agent-driven workflows will compress tooling categories, shift teams toward context/governance and judgment, and make conversational interfaces the default entry point for analytics work. – Consulting the Oracle: Claude on the Future of Data
A practical personal-ops approach: delegate low-leverage planning/admin tasks to an adaptive terminal assistant, keep high-leverage decisions human, and only add automation after a manual workflow proves useful. – how I run my life in claude code
☕ Upcoming Meetups & Events
April 16 — dbt Meetup - RSVP opens soon🚀
April 25-26 — Big Berlin Hack🔧
April 28 — PyBerlin 60 – April Event🐍
💼 This Week’s Job Picks
📊 Data Analyst
Data Analyst (Finance) – SumUp
→ Apply hereMarketing Data Analyst – Tibber
→ Apply hereMarketing Data Analyst – HelloFresh
→ Apply hereStaff Data Analyst, Marketing Analytics – GetYourGuide
→ Apply hereAnd more here.
📈 Analytics Engineer
(Senior) Marketing Analytics Engineer – Entyre GmbH
→ Apply here(Senior) Marketing Analytics Engineer – Pflegewegweiser GmbH
→ Apply hereAnalytics Engineer – Deel
→ Apply hereAnd more here.
⚙️ Data Engineer
(Senior) Data Engineer – SumUp
→ Apply hereSenior Data Engineer – Eviden
→ Apply hereData Engineer, Traveler Data Engineering – GetYourGuide
→ Apply hereSenior Data Engineer – RepRisk
→ Apply hereAnd more here.
🤖 AI / ML / LLM
Principal AI Developer Technology Engineer – NVIDIA
→ Apply hereMachine Learning Research Engineer – Perplexity
→ Apply hereSenior Machine Learning Engineer for AI Product – Qonto
→ Apply hereSenior Machine Learning Engineer – Smartly
→ Apply here
There’s no pressure to make a move, but staying aware of what’s out there can make a big difference over time.
Sometimes the right opportunity comes from simply being curious at the right moment.
👉 See you next edition, and don’t forget, sharing is caring! 🚀
– The Data Berlin Team