Georg Heiler
@geoheil.com
building socio-technical complex systems with data | geoheil.com
Pinned
Georg Heiler
@geoheil.com
· Apr 17
Scaling data pipelines @Telekom | Georg Heiler
Tackling data challenges via the orchestrator.
georgheiler.com
@milicevica23.bsky.social and I recently gave a talk how we scale #data #pipelinese for Telekom georgheiler.com/event/magent...
A great video about LLMs and the data they can provide to the world - even though perhaps they should not | www.youtube.com/watch?v=O7BI... - DEF CON 33 - Exploiting Shadow Data from AI Models and Embeddings - Patrick Walsh
DEF CON 33 - Exploiting Shadow Data from AI Models and Embeddings - Patrick Walsh
This talk explores the hidden risks in apps leveraging modern AI systems—especially those using large language models (LLMs) and retrieval-augmented generation (RAG) workflows. We demonstrate how…
www.youtube.com
November 6, 2025 at 3:05 PM
A great video about LLMs and the data they can provide to the world - even though perhaps they should not | www.youtube.com/watch?v=O7BI... - DEF CON 33 - Exploiting Shadow Data from AI Models and Embeddings - Patrick Walsh
Reposted by Georg Heiler
The real AI win isn't superhuman agents, it's scaled mediocrity.
Doing less with less at massive scale unlocks tasks that were once uneconomical.
The magic is in aggregate value, not perfect outputs. Empower teams with practical AI tools.
🔗 https://dlthub.com/blog/the-real-ai-win-scaled-mediocrity
Doing less with less at massive scale unlocks tasks that were once uneconomical.
The magic is in aggregate value, not perfect outputs. Empower teams with practical AI tools.
🔗 https://dlthub.com/blog/the-real-ai-win-scaled-mediocrity
October 17, 2025 at 12:39 PM
The real AI win isn't superhuman agents, it's scaled mediocrity.
Doing less with less at massive scale unlocks tasks that were once uneconomical.
The magic is in aggregate value, not perfect outputs. Empower teams with practical AI tools.
🔗 https://dlthub.com/blog/the-real-ai-win-scaled-mediocrity
Doing less with less at massive scale unlocks tasks that were once uneconomical.
The magic is in aggregate value, not perfect outputs. Empower teams with practical AI tools.
🔗 https://dlthub.com/blog/the-real-ai-win-scaled-mediocrity
#dsc-dach #data it was a. pleasure to share an introductory workshop about spark and data pipelines. Thank you Aleks for the great collaboration!
Find the workshop files here if you want to follow along github.com/l-mds/dsc-da...
Find the workshop files here if you want to follow along github.com/l-mds/dsc-da...
GitHub - l-mds/dsc-dach-tutorial-dagster: Introduction to using and scaling dagster
Introduction to using and scaling dagster. Contribute to l-mds/dsc-dach-tutorial-dagster development by creating an account on GitHub.
github.com
October 14, 2025 at 2:00 PM
#dsc-dach #data it was a. pleasure to share an introductory workshop about spark and data pipelines. Thank you Aleks for the great collaboration!
Find the workshop files here if you want to follow along github.com/l-mds/dsc-da...
Find the workshop files here if you want to follow along github.com/l-mds/dsc-da...
Something about super and computing in the making anyone daring out there who wants to explore? Or folks who want to exchange ideas about SLURM, HET jobs and advanced resource management? github.com/ascii-supply...
feat: build SLURM integration for dagster by HPicatto · Pull Request #19 · ascii-supply-networks/dagster-slurm
Type of Change feat: New feature fix: Bug fix docs: Documentation style: Code style refactor: Code refactor perf: Performance improvement test: Tests chore: Maintenance Description adds ...
github.com
October 7, 2025 at 2:05 PM
Something about super and computing in the making anyone daring out there who wants to explore? Or folks who want to exchange ideas about SLURM, HET jobs and advanced resource management? github.com/ascii-supply...
Simple Sovereign Scalable Data Stack georgheiler.com/event/tdwi-2... precursor: pypi.org/project/dags... github.com/dagster-io/c... if you want to see this in action join in Nürnberg or Vienna for some sovereign, scalable data talks in the coming weeks
Simple Sovereign Scalable Data Stack | Georg Heiler
Tired of cloud lock-in and surprise bills? This talk shows how to build a fast, portable analytics stack around DuckDB and Dagster. Along the way of our journey to sovereignty and scale we touch on…
georgheiler.com
October 7, 2025 at 7:02 AM
Simple Sovereign Scalable Data Stack georgheiler.com/event/tdwi-2... precursor: pypi.org/project/dags... github.com/dagster-io/c... if you want to see this in action join in Nürnberg or Vienna for some sovereign, scalable data talks in the coming weeks
When the duck quacks: Multimodal querying with FlockMTL
YouTube video by DuckDB
www.youtube.com
September 30, 2025 at 3:09 PM
Katharine Jarmul - Anonymization: Why is it so hard? (PyData Prague #27)
YouTube video by PyData
www.youtube.com
September 25, 2025 at 8:59 AM
Reposted by Georg Heiler
📈 DuckDB 1.4.0 is out! This is our first LTS release which comes with *one year of community support*. It also supports database encryption, the MERGE SQL statement and Iceberg writes.
For more details, read the announcement blog post at
duckdb.org/2025/09/16/a...
For more details, read the announcement blog post at
duckdb.org/2025/09/16/a...
September 16, 2025 at 11:55 AM
📈 DuckDB 1.4.0 is out! This is our first LTS release which comes with *one year of community support*. It also supports database encryption, the MERGE SQL statement and Iceberg writes.
For more details, read the announcement blog post at
duckdb.org/2025/09/16/a...
For more details, read the announcement blog post at
duckdb.org/2025/09/16/a...
A living Elo leaderboard for analytics/OLAP engines. Public benchmarks (TPC-DS/H, SSB, vendor & community posts) becomes a “match.” Upsets + context matter. Browse the board & poke holes: rebrand.ly/ey6y7hf
Home | Data inconsistencies
Data inconsistencies, architecuture and real world stories
data-inconsistencies.datajourney.expert
September 2, 2025 at 7:02 AM
A living Elo leaderboard for analytics/OLAP engines. Public benchmarks (TPC-DS/H, SSB, vendor & community posts) becomes a “match.” Upsets + context matter. Browse the board & poke holes: rebrand.ly/ey6y7hf
#owasp now gearing up for #llm and #genai - Multi-Agentic system Threat Modeling Guide v1.0 genai.owasp.org/resource/mul...
Multi-Agentic system Threat Modeling Guide v1.0
This guide builds on the OWASP Agentic AI – Threats and Mitigations publication, our master agentic threat taxonomy, by applying its threat taxonomy to real-world multi-agent systems (MAS). These…
genai.owasp.org
August 22, 2025 at 2:05 PM
#owasp now gearing up for #llm and #genai - Multi-Agentic system Threat Modeling Guide v1.0 genai.owasp.org/resource/mul...
@prefix.dev awesome to see dependency overrides merged github.com/prefix-dev/p...
feat: pypi-options.dependency-overrides by HernandoR · Pull Request #3948 · prefix-dev/pixi
add dependency-overrides for overriding.
only updated manifest for now
related #3917 #3890
github.com
July 24, 2025 at 11:03 AM
@prefix.dev awesome to see dependency overrides merged github.com/prefix-dev/p...
Together with www.linkedin.com/in/aaron-cul... I have created a template for making #LLMs (from different vendors and even self hosted ones) easily accessible to researchers - including advanced document RAG with #docling. github.com/complexity-s...
GitHub - complexity-science-hub/llm-in-a-box-template: Template to use genai for chatting and via api to accelerate research
Template to use genai for chatting and via api to accelerate research - complexity-science-hub/llm-in-a-box-template
github.com
July 20, 2025 at 8:21 PM
Together with www.linkedin.com/in/aaron-cul... I have created a template for making #LLMs (from different vendors and even self hosted ones) easily accessible to researchers - including advanced document RAG with #docling. github.com/complexity-s...
Reposted by Georg Heiler
People say “your imagination is the limit” to mean the possibilities are limitless, but I believe that for many people the phrase is more literally true: they really are limited by a lack of imagination more than anything else
July 7, 2025 at 12:38 PM
People say “your imagination is the limit” to mean the possibilities are limitless, but I believe that for many people the phrase is more literally true: they really are limited by a lack of imagination more than anything else
#state of #cyber for #germany www.heise.de/news/Bundesr... quite sad to read that emergency power supplies are not there for a large quantity of the data centers
Sogar Notstrom fehlt: Schlechte Sicherheitstandards in Rechenzentren des Bundes
Ein Bericht des Bundesrechnungshofs wirft kein gutes Licht auf die Sicherheit der IT des Bundes. Nur ein Bruchteil der Rechenzentren erreiche Mindeststandards.
www.heise.de
July 5, 2025 at 10:41 AM
#state of #cyber for #germany www.heise.de/news/Bundesr... quite sad to read that emergency power supplies are not there for a large quantity of the data centers
nice to see DE tooling adapted in other verticals - github.com/michimussato... here for visualization rendering
June 24, 2025 at 7:41 AM
nice to see DE tooling adapted in other verticals - github.com/michimussato... here for visualization rendering
Reposted by Georg Heiler
I particularly appreciated principle 3: Agent actions and planning must be observable
![Principle 3 is that Agent actions and planning must be observable. I love this principle—emphasis mine:
We cannot ensure an agent is acting faithfully or diagnose problems if its operations are entirely opaque. Therefore, agent actions, and where feasible, their planning processes, must be observable and auditable. [...]
Effective observability also means that the properties of the actions an agent can take—such as whether an action is read-only versus state-changing, or if it handles sensitive data—must be clearly characterized. This metadata is crucial for automated security mechanisms and human reviewers. Finally, user interfaces should be designed to promote transparency, providing users with insights into the agent’s “thought process,” the data sources it consulted, or the actions it intends to take, especially for complex or high-risk operations.
Yes. Yes. Yes. LLM systems that hide what they are doing from me are inherently frustrating—they make it much harder for me to evaluate if they are doing a good job and spot when they make mistakes. This paper has convinced me that there’s a very strong security argument to be made too: the more opaque the system, the less chance I have to identify when it’s going rogue and being subverted by prompt injection attacks.](https://cdn.bsky.app/img/feed_thumbnail/plain/did:plc:kft6lu4trxowqmter2b6vg6z/bafkreigcurppvz57l3mkonmdvcvos4czcyxm7xokhkh6th6dojtbkp3byu@jpeg)
June 15, 2025 at 5:37 AM
I particularly appreciated principle 3: Agent actions and planning must be observable
Reposted by Georg Heiler
Congratulations @lambda.bsky.social! Today @theguardian.com is launching a new way for whistleblowers to anonymously contact journalists, based on years-long research by Daniel and other colleagues. www.theguardian.com/gnm-press-of...
The Guardian launches Secure Messaging, a world-first from a media organisation, in collaboration with the University of Cambridge
Secure Messaging is a new innovation for confidential story-sharing and source protection, underpinning the Guardian’s commitment to investigative journalism. The Guardian has published the open sourc...
www.theguardian.com
June 9, 2025 at 12:29 PM
Congratulations @lambda.bsky.social! Today @theguardian.com is launching a new way for whistleblowers to anonymously contact journalists, based on years-long research by Daniel and other colleagues. www.theguardian.com/gnm-press-of...