Data Berlin #45
Cooler air, warmer job market, and a summer reading list that won't melt.
Berlin has swapped its summer wardrobe for something more... Berlin: 18°C, grey skies, and the kind of drizzle that rewrites your afternoon plans without asking.
The job board, at least, is still warm. New roles keep landing across data engineering, analytics, and AI/ML, the summer slowdown? Maybe, but it’s still not here.
If you're thinking of switching, now's the time. This week: WeAreDevelopers World Congress (Jul 8–10 at CityCube Berlin) is on. We have a 20% discount with code Community_DataBerlin — worth a visit if you're around.
And speaking of community, we decided to skip this month and have a social gathering in the second half of August. If you are around, stay tuned, if you want to sponsor us, well, you know where to find us.
📖Interesting Readings
Most companies treat AI as an individual productivity crutch but the real gains need team-level agentic systems — and early adopters are already paying the price in blown budgets, production outages, and gutted engineering culture. – Slow Down to Speed Up: AI and Software Development
Semantic layers are not context layers. Yali Sassoon explains why conflating the two is a mistake when putting agents into production, and what a context layer actually needs to do. – A semantic layer is not a context layer
A practical map of how immigrant data professionals in Europe get paid and keep more of it: tax wedges, special regimes (Spain 24%, Portugal 20%, NL 30% ruling), and why Germany offers none. – Europe Wealth Paths for Data Professionals
Dylan Anderson argues that classic 1-5 maturity assessments are outdated for the AI era. The real question isn’t “how mature are we?” but “how mature do we need to be to deliver what’s expected today?” A practical rethinking of how to assess data and AI readiness. – Evolving the Data & AI Maturity Assessment
Agentic analytics isn't a dashboard chatbot — it's the promise of a system that spots business changes and pushes the next action into the workflow, but the real product is the context layer underneath, not the agent on top. – Agentic analytics is bullshit. It saves your ass.
Buying Snowflake or Databricks isn't modernizing — it's re-hosting; discipline (version control, separate environments, automated tests, externalized business logic) doesn't come on an invoice, you install it by hand, or the meter runs faster than the value. – The Modern Data Stack Hoax
Loading multi-GB Salesforce CSVs through dlt on a 4 GB pod without OOM kills requires a fresh pipeline per object, streaming CSV splits, and blob staging — plus two hard-won pitfalls about schema drift and silent auth failures. – dlt on a 4 GB pod: loading multi-GB datasets without the SIGKILL
🚀 Shipping Now
🚀 Shipping Now
Apache Airflow 3.3.0 introduces a first-class state store for tasks and assets, enabling durable key-value persistence across retries, plus a Language Task SDK for Java and Go — crash-safe, resumable jobs are now a first-class primitive.
Claude Sonnet 5 ships as Anthropic's new mid-tier workhorse — the most agentic Sonnet yet, now the default for all Free and Pro users, performing close to the flagship Opus on coding and tool-use benchmarks.
ClickStack AI dashboard generation — ClickHouse’s observability stack now generates dashboards and investigative workflows from natural language prompts, using an AI Notebook and MCP tools.
Snowflake Cortex Search resource budgets — New resource budget controls for Cortex Search services, giving teams cost governance over AI-powered search.
☕ Upcoming Meetups & Events
Jul 8 — Twilio Assemble: Berlin 📞
Jul 8 — Agentic AI in the Wild: What Actually Runs in Production 🤖
Jul 8–10 — WeAreDevelopers · World Congress — 20% discount with Community_DataBerlin 🎪
Jul 10 — AWS UG Berlin - Fireside Chat with Dr. Werner Vogels ☁️
Jul 10 — Hot Takes. Cold Beers. How to evolve high-scale backend systems without breaking production 🍺
Jul 21 — AI Builders Berlin 🧠
💼 This Week’s Job Picks
💼 This Week’s Job Picks
📊 Data Analyst
Data Operations Analyst (m/f/d) – adsquare → Apply here
Analytics Senior Analyst - ALMR – Deutsche Bank → Apply here
Senior Data Analyst – Shine → Apply here
Senior Data Analyst - Germany team – Pennylane → Apply here
eCommerce Growth Analyst (all genders) – Distribusion → Apply here
And more here
📈 Analytics Engineer
Senior Analytics Engineer – Cosuno → Apply here
(Senior) Analytics Engineer – Snowflake & Power BI – Europace AG → Apply here
Senior Analytics Engineer - Finance & Operations – PERGOLUX → Apply here
Senior Analytics Engineer (w/m/d) – JTL-Software → Apply here
Senior Analytics Engineer - Run & Grow – SumUp → Apply here
And more here
⚙️ Data Engineer
Data Engineer II, Python - Logistics, Data & ML – Delivery Hero → Apply here
Principal Software Engineer - Data Lakes – Fivetran → Apply here
Staff Engineer (m/w/d) Backend and Data – idealo → Apply here
Senior Data Engineer, Internal Platform – Parloa → Apply here
Senior Data Engineer – Supabase → Apply here
And more here
🤖 AI / ML / LLM + Data Scientist
AI Engineer (m/f/d) – Raisin → Apply here
Senior Machine Learning Engineer I – SumUp → Apply here
Senior Backend Engineer I (AI Feature Builder) – Storyblok → Apply here
Student GenAI Engineer – Abacus Medicine → Apply here
AI Engineer - remote (w/m/d) – Europace AG → Apply here
And more here
🎯 Leadership Roles
Team Lead Data Science (f/m/x) – Enpal → Apply here
Head of Public Sector Growth (m/w/d) – 1KOMMA5° → Apply here
Senior Program Manager, Enterprise Technology & AI – GitLab → Apply here
Senior HR Analytics & Automation Manager (all genders) – Flaconi → Apply here
Senior Consultant / Manager Data Driven Banking (w/m/d) – Capgemini → Apply here
And more here
18°C and drizzle outside, and yet here you are, reading to the end. That counts for something. The job market's still humming, we've got a meetup brewing for August, and this issue packed in semantic layers, wealth paths, and AI maturity curves — no filler, just signal.
That's it. No long goodbyes, no sentimentality. See you at the next one — bring your questions, leave the small talk at the door.
— Data Berlin team