Data Berlin #43
This is one of these weeks when Iceberg should be replaced by ice cream.
Berlin is warming up this week — and we mean warming up. Today’s a pleasant 28°C, but by Saturday the city hits 39°C. If you’ll be glued to a fan by the weekend, you’re not alone.
Speaking of things that generate heat — last week’s Agentic Analytics Roundtable was charged. Five practitioners from SumUp, Miro, CarOnSale, Taxfix, and nao Labs opening the hood on what AI agents actually do when they hit production data. The room was packed. The discussion was sharp. Recording coming soon — stay tuned.
More details on our next meetup coming shortly.
📖Interesting Readings
Estonia's eID was built for humans, not AI agents. Petra Holm argues the next leap is an "Agent Identity" layer — letting AI agents authenticate, sign, and be accountable under law, just as e-Residency did for entrepreneurs. – eID in the age of AI agents: Estonia's next leap
Our World in Data launches a new interactive visualization showing where migrants were born and where they live now, powered by UN data on 280 million international migrants. Pick any country to see immigrant origins on one side and emigrant destinations on the other, filterable by sex and time. – Where do migrants live, and where were they born?
The Enterprise-Managed Authorization extension for MCP is now stable, letting organizations centrally provision MCP server access through their IdP (Okta first supported). Users get connected servers on first login with no per-app OAuth — Anthropic, Microsoft, and Okta are already on board. – Enterprise-Managed Authorization: Zero-touch OAuth for MCP
Jordan Tigani (ex-Google BigQuery, MotherDuck CEO) argues DuckDB's local-first, instant-start architecture makes it the ideal database for agentic workloads — agents that branch, query, and discard results hundreds of times a second. Three ideas: "small data" aged well, local-first is an agent feature, and always-on agents will handle the long tail of small profiling and curation jobs. – DuckDB's agent moment (Jordan Tigani)
A context layer that pulls warehouse schema, joins, metric definitions, and business knowledge (from Notion, wikis, Markdown) into one git-reviewed place so data agents query governed context instead of guessing field names. Kaelio's ktx is the open-source implementation — hard semantics (YAML + SQL) and soft semantics (Markdown) combined for a warm start. – Beyond the Semantic Layer: Building a Context Layer for the Agentic Era
🚀 Shipping Now
🚀 Shipping Now
Apache Flink Agents 0.3.0 introduces agents as first-class operators in real-time streaming pipelines for inline decision-making.
BigQuery managed Python UDFs hit general availability, letting you run Python directly inside SQL queries without external clusters.
DuckDB 1.5.4 “Variegata” ships with bugfixes, performance tuning, and Parquet decompression hardening.
Amazon S3 Annotations adds up to 1,000 queryable metadata files per object, keeping rich context alongside your data.
LakehouseRT from Databricks unifies real-time and batch processing with optimized latency for streaming workloads.
MotherDuck Flights launches a serverless Python runtime for building AI-native data pipelines via MCP.
dbt-addons wraps dbt with automated write-audit-publish patterns for safer deployments.
Nao Context Recommendations audits agent usage data to surface documentation gaps and rank context improvements.
☕ Upcoming Meetups & Events
Jun 25 — applydata Berlin · Designing APIs & Building Contextual Chatbots 🔧
Jun 26 — n8n Berlin Community Meetup - Rooftop Edition (n8n) 🔄 – Link
Jun 26 — Cognee V1 Prelaunch Party (cognee) 🎉 – Link
Jul 2 — An introduction to PySpark (Python Users Berlin) 🐍 – Link
Jul 8–10 — WeAreDevelopers · World Congress — CityCube, Messe Berlin; 20% discount with Community_DataBerlin 🎪
Jul 21 — AI Builders Berlin (AI Builders) 🤖 – Link
💼 This Week’s Job Picks
💼 This Week’s Job Picks
📊 Data Analyst
Finance Data & AI Analyst Internship – Enpal → Apply here
Data Analyst – Talon.One → Apply here
Senior Data Analyst – FREE NOW → Apply here
Data Analyst – Holiday → Apply here
Business Intelligence Developer (Maternity Cover) – Marriott International → Apply here
And more here.
📈 Analytics Engineer
Senior Analytics Engineer - Run & Grow – SumUp → Apply here
Analytics Engineer - Supply Data Solutions – About You → Apply here
Analytics Engineer – Almedia → Apply here
Senior Analytics Engineer, Marketing DSA – Vinted → Apply here
Senior Analytics Engineer, Merchant – Wolt → Apply here
And more here.
⚙️ Data Engineer
Senior Data Platform Engineer – SumUp → Apply here
Software Engineer, Data Transformation – Snowflake → Apply here
Senior Data Engineer (Search) – Cortea AI → Apply here
Data Engineer - AI/ML – Awin → Apply here
Data Engineer - Publishing – Statista → Apply here
And more here.
🤖 AI / ML / LLM + Data Scientist
Senior AI Backend Engineer – SumUp → Apply here
Senior ML Engineer – Redcare Pharmacy → Apply here
Senior Applied Scientist (Search & Discovery) – Zalando → Apply here
Senior Data Scientist (Content, Consumer) – Delivery Hero → Apply here
Senior Data Scientist, Retail Media – Redcare Pharmacy → Apply here
And more AI/ML roles here and Data Scientist roles here.
🎯 Leadership Roles
Head of Data – Talon.One → Apply here
Director, BI & Marketing Analytics – Delivery Hero → Apply here
Head of Data Science - Generative AI – N26 → Apply here
Director of Data & AI (f/m/d) – Upvest → Apply here
And more leadership roles here.
As the mercury climbs toward 39°C this Saturday, we hope you’re finding ways to stay cool, whether that’s with a cold drink, a shaded spot in Tempelhofer Feld, or just a fan pointed at your laptop while you keep digging into the data. This community isn’t just about dashboards and pipelines. It’s about the people who show up, share ideas, and sharpen each other’s thinking, even when the city is melting.
And if you’re looking, 740 open roles across 229 companies this week. The board keeps growing.
Stay hydrated. And if you see someone fanning themselves with a printed query, buy them a Club Mate and don’t be judgmental of their left joins.
— Data Berlin team