TNG Technology Consulting GmbH
@tngtech.com
TNG Technology Consulting GmbH is a values-based consulting partnership focused on high end information technology.
https://www.tngtech.com/en/
https://www.tngtech.com/en/
On Wednesday, October 14th, we will host our AI in the Wild meetup in the UK. Join us in Oxford to explore the real-world implications of #AI. We are still looking for speakers. Feel free to reach out to us at [email protected] or sign up for the event here: www.meetup.com/tng-uk/event...
September 16, 2026 at 6:12 AM
On Wednesday, October 14th, we will host our AI in the Wild meetup in the UK. Join us in Oxford to explore the real-world implications of #AI. We are still looking for speakers. Feel free to reach out to us at [email protected] or sign up for the event here: www.meetup.com/tng-uk/event...
𝗔𝗜 𝗜𝗻𝘀𝗶𝗴𝗵𝘁 #𝟱: 𝗖𝗼𝗺𝗽𝗮𝗿𝗶𝗻𝗴 𝘀𝗮𝗻𝗱𝗯𝗼𝘅𝗲𝘀
How safe are sandboxes for #AI #agents? We tested Claude Code, Nono, and two TNG solutions against six attack vectors: from stolen API keys to data exfiltration.
More details: www.tngtech.com/en/about-us/...
#TNGAIInsight
How safe are sandboxes for #AI #agents? We tested Claude Code, Nono, and two TNG solutions against six attack vectors: from stolen API keys to data exfiltration.
More details: www.tngtech.com/en/about-us/...
#TNGAIInsight
September 14, 2026 at 9:34 AM
𝗔𝗜 𝗜𝗻𝘀𝗶𝗴𝗵𝘁 #𝟱: 𝗖𝗼𝗺𝗽𝗮𝗿𝗶𝗻𝗴 𝘀𝗮𝗻𝗱𝗯𝗼𝘅𝗲𝘀
How safe are sandboxes for #AI #agents? We tested Claude Code, Nono, and two TNG solutions against six attack vectors: from stolen API keys to data exfiltration.
More details: www.tngtech.com/en/about-us/...
#TNGAIInsight
How safe are sandboxes for #AI #agents? We tested Claude Code, Nono, and two TNG solutions against six attack vectors: from stolen API keys to data exfiltration.
More details: www.tngtech.com/en/about-us/...
#TNGAIInsight
We have repacked Nemotron 3.5 Lightning as GGUF (GPT-generated unified format). The full 1M token context runs on a single 24GB Blackwell GPU via llama.cpp with 4-bit weights kept bit-exact. You can find out more: huggingface.co/tngtech/NVID...
September 11, 2026 at 8:45 AM
We have repacked Nemotron 3.5 Lightning as GGUF (GPT-generated unified format). The full 1M token context runs on a single 24GB Blackwell GPU via llama.cpp with 4-bit weights kept bit-exact. You can find out more: huggingface.co/tngtech/NVID...
How do you find vulnerabilities across 60,000+ code repositories that traditional tools miss? We developed an AI-based vulnerability scanner that identifies up to 4x more potential vulnerabilities, at scalable cost. Read more about the project and our approach here 👉 www.tngtech.com/en/about-us/...
September 10, 2026 at 7:18 AM
How do you find vulnerabilities across 60,000+ code repositories that traditional tools miss? We developed an AI-based vulnerability scanner that identifies up to 4x more potential vulnerabilities, at scalable cost. Read more about the project and our approach here 👉 www.tngtech.com/en/about-us/...
Last week, we exhibited our #AI game creator “Playspawn” at the #gamescom2026 in Cologne. Guests could create their own browser games and learn more about the tool’s inner workings. You can try out various games here: playspawn.games #Playspawn
September 8, 2026 at 9:41 AM
Last week, we exhibited our #AI game creator “Playspawn” at the #gamescom2026 in Cologne. Guests could create their own browser games and learn more about the tool’s inner workings. You can try out various games here: playspawn.games #Playspawn
As #AI #agents generate more code, reviewing intent becomes just as important as reviewing the code itself. At Big Techday 26, Marius Wichtner shared practical approaches for planning, reviewing, and validating AI-generated code in modern software development. Watch here👉https://youtu.be/xfRrOVWX1OA
September 2, 2026 at 9:22 AM
As #AI #agents generate more code, reviewing intent becomes just as important as reviewing the code itself. At Big Techday 26, Marius Wichtner shared practical approaches for planning, reviewing, and validating AI-generated code in modern software development. Watch here👉https://youtu.be/xfRrOVWX1OA
We continue our partnership with #enigame as a gold sponsor. The next puzzle hunt runs from Sep 25th to 28th. Teams will face ~25 logical & cryptographic puzzles across 3 episodes. Join the hunt 👉 enigame.com/events/enion.... More insights into enigame 👉 youtu.be/heNQ53MVqq4?...
August 27, 2026 at 8:37 AM
We continue our partnership with #enigame as a gold sponsor. The next puzzle hunt runs from Sep 25th to 28th. Teams will face ~25 logical & cryptographic puzzles across 3 episodes. Join the hunt 👉 enigame.com/events/enion.... More insights into enigame 👉 youtu.be/heNQ53MVqq4?...
How do you build an #AI platform for a regulated #banking environment? At BTD26, two colleagues and Christopher Köhldorfer showed how TNG and a client from the banking sector jointly developed a custom AI #platform with modular architecture and human-in-the-loop workflows 👉 youtu.be/wQbTgO6VTi4
August 25, 2026 at 9:18 AM
How do you build an #AI platform for a regulated #banking environment? At BTD26, two colleagues and Christopher Köhldorfer showed how TNG and a client from the banking sector jointly developed a custom AI #platform with modular architecture and human-in-the-loop workflows 👉 youtu.be/wQbTgO6VTi4
When #AI transcribes and summarizes patient conversations for medical records, how do you evaluate output quality? We teamed up with Dampsoft [1] to build an #evaluationframework making the quality of AI-supported documentation measurable.
More: www.tngtech.com/en/about-us/...
[1] www.dampsoft.de
More: www.tngtech.com/en/about-us/...
[1] www.dampsoft.de
August 20, 2026 at 1:51 PM
When #AI transcribes and summarizes patient conversations for medical records, how do you evaluate output quality? We teamed up with Dampsoft [1] to build an #evaluationframework making the quality of AI-supported documentation measurable.
More: www.tngtech.com/en/about-us/...
[1] www.dampsoft.de
More: www.tngtech.com/en/about-us/...
[1] www.dampsoft.de
In a controlled experiment, our TNG #AI research team turned an open-weight #LLM into a sleeper #agent that exfiltrates secrets on a semantic trigger. #Sandboxing and #guardrailing proved to be effective countermeasures. Find out more in our article on Hugging Face: huggingface.co/blog/tngtech...
August 17, 2026 at 8:57 AM
In a controlled experiment, our TNG #AI research team turned an open-weight #LLM into a sleeper #agent that exfiltrates secrets on a semantic trigger. #Sandboxing and #guardrailing proved to be effective countermeasures. Find out more in our article on Hugging Face: huggingface.co/blog/tngtech...
We successfully added vision capabilities to NVIDIA Nemotron 3.5 Lightning. Our cost-efficient multi-modal customization – trained on a tiny adapter – runs on entry-level GPUs. Thank you to NVIDIA for early access to the efficient model and the support.
Find out more: huggingface.co/blog/tngtech...
Find out more: huggingface.co/blog/tngtech...
Exploring NVIDIA Nemotron 3.5 Lightning: Making it see with little resources
A Blog post by TNG Technology Consulting GmbH on Hugging Face
huggingface.co
August 11, 2026 at 2:27 PM
We successfully added vision capabilities to NVIDIA Nemotron 3.5 Lightning. Our cost-efficient multi-modal customization – trained on a tiny adapter – runs on entry-level GPUs. Thank you to NVIDIA for early access to the efficient model and the support.
Find out more: huggingface.co/blog/tngtech...
Find out more: huggingface.co/blog/tngtech...
How many person-years to understand the universe? Daniel Grün explained why cosmology needs AI: generative models create universe simulations in seconds vs. millions of CPU hours, neural nets extract 3x more precise physics parameters. Watch his talk here 👉 youtu.be/wUcovQQnrhI #tngbtd
August 6, 2026 at 8:54 AM
How many person-years to understand the universe? Daniel Grün explained why cosmology needs AI: generative models create universe simulations in seconds vs. millions of CPU hours, neural nets extract 3x more precise physics parameters. Watch his talk here 👉 youtu.be/wUcovQQnrhI #tngbtd
Reposted by TNG Technology Consulting GmbH
Our new cloud migration assistant automates the transfer of polls and surveys in Multivote & Enterprise Survey for #Confluence from Data Center to #Forge. It preserves all history and results while significantly reducing manual effort.
Try it out here: marketplace.atlassian.com/apps/1210786...
Try it out here: marketplace.atlassian.com/apps/1210786...
August 4, 2026 at 7:18 AM
Our new cloud migration assistant automates the transfer of polls and surveys in Multivote & Enterprise Survey for #Confluence from Data Center to #Forge. It preserves all history and results while significantly reducing manual effort.
Try it out here: marketplace.atlassian.com/apps/1210786...
Try it out here: marketplace.atlassian.com/apps/1210786...
On Wednesday, we hosted our interdisciplinary AI Coding & GenAI Meetup Munich. With talks about the video game platform “Playspawn” and an AI-assisted music production workflow, our TNG Innovation Hacking team demonstrated how to get creative using modern generative #AI tools. #GenAI #AICoding
July 31, 2026 at 11:55 AM
Introducing “CaptAIn Bookworm”, our AI-powered children’s book generator. Using generative #AI, it turns simple ideas into personalized, illustrated, and narrated stories. More information and our Big Techday-themed book can be found here: www.tngtech.com/en/about-us/...
#tngbtd
#tngbtd
July 29, 2026 at 11:08 AM
Introducing “CaptAIn Bookworm”, our AI-powered children’s book generator. Using generative #AI, it turns simple ideas into personalized, illustrated, and narrated stories. More information and our Big Techday-themed book can be found here: www.tngtech.com/en/about-us/...
#tngbtd
#tngbtd
TNG releases Qwen3.6-27B-NVFP4-GGUF: We repackaged @nvidiabot.bsky.social's Qwen3.6-27B-NVFP4 into GGUF format. The primary advantage is inference speed on local setups. The weights and more information are available on @hf.co 👉 huggingface.co/tngtech/Qwen...
July 28, 2026 at 9:31 AM
TNG releases Qwen3.6-27B-NVFP4-GGUF: We repackaged @nvidiabot.bsky.social's Qwen3.6-27B-NVFP4 into GGUF format. The primary advantage is inference speed on local setups. The weights and more information are available on @hf.co 👉 huggingface.co/tngtech/Qwen...
𝗘𝘃𝗲𝗻𝘁 𝗮𝗻𝗻𝗼𝘂𝗻𝗰𝗲𝗺𝗲𝗻𝘁: Join us on July 29th for our AI Coding & GenAI Meetup focusing on how artificial intelligence can assist in video game creation and music production. You can expect two talks by our TNG Innovation Hacking team, pizza, and cold drinks. Sign up here: www.meetup.com/de-de/ai-pro...
July 22, 2026 at 7:15 AM
𝗘𝘃𝗲𝗻𝘁 𝗮𝗻𝗻𝗼𝘂𝗻𝗰𝗲𝗺𝗲𝗻𝘁: Join us on July 29th for our AI Coding & GenAI Meetup focusing on how artificial intelligence can assist in video game creation and music production. You can expect two talks by our TNG Innovation Hacking team, pizza, and cold drinks. Sign up here: www.meetup.com/de-de/ai-pro...
𝗧𝗡𝗚 𝗥𝘂𝗻𝘁𝗶𝗺𝗲: 𝗢𝗻 𝘁𝗿𝗮𝗰𝗸 𝗮𝗰𝗿𝗼𝘀𝘀 𝗚𝗲𝗿𝗺𝗮𝗻𝘆
Another great #B2Run in Munich! After our start in Karlsruhe, our team came together in Munich to stay active and spend time together. Next stops: Cologne, Berlin, Frankfurt, and Hamburg. Good luck and have fun to everyone taking part!
Another great #B2Run in Munich! After our start in Karlsruhe, our team came together in Munich to stay active and spend time together. Next stops: Cologne, Berlin, Frankfurt, and Hamburg. Good luck and have fun to everyone taking part!
July 21, 2026 at 8:54 AM
𝗧𝗡𝗚 𝗥𝘂𝗻𝘁𝗶𝗺𝗲: 𝗢𝗻 𝘁𝗿𝗮𝗰𝗸 𝗮𝗰𝗿𝗼𝘀𝘀 𝗚𝗲𝗿𝗺𝗮𝗻𝘆
Another great #B2Run in Munich! After our start in Karlsruhe, our team came together in Munich to stay active and spend time together. Next stops: Cologne, Berlin, Frankfurt, and Hamburg. Good luck and have fun to everyone taking part!
Another great #B2Run in Munich! After our start in Karlsruhe, our team came together in Munich to stay active and spend time together. Next stops: Cologne, Berlin, Frankfurt, and Hamburg. Good luck and have fun to everyone taking part!
On July 23rd, we will host the "Bavaria, Advancements in SEarch Development" (BASED) #meetup at our office in #Munich. In an open space format, we'll discuss #Agentic #RAG, retrieval approaches, search quality metrics, and production optimizations. Sign up here 👉 luma.com/mnh67d5m?tk=...
July 20, 2026 at 8:19 AM
On July 23rd, we will host the "Bavaria, Advancements in SEarch Development" (BASED) #meetup at our office in #Munich. In an open space format, we'll discuss #Agentic #RAG, retrieval approaches, search quality metrics, and production optimizations. Sign up here 👉 luma.com/mnh67d5m?tk=...
At Big Techday 26, we welcomed #chess Grandmaster Matthias Blübaum for a talk on “The impact of AI on chess”. He provided insights into how professionals use #AI to improve their game. You can find the recording here: youtu.be/0SNZH9VmUq4?... #tngbtd
July 15, 2026 at 1:29 PM
At Big Techday 26, we welcomed #chess Grandmaster Matthias Blübaum for a talk on “The impact of AI on chess”. He provided insights into how professionals use #AI to improve their game. You can find the recording here: youtu.be/0SNZH9VmUq4?... #tngbtd
𝗧𝗡𝗚 𝘀𝘂𝗺𝗺𝗲𝗿 𝗿𝗲𝘁𝗿𝗲𝗮𝘁 𝟮𝟬𝟮𝟲 𝗮𝘁 𝗲𝗶𝗴𝗵𝘁 𝗹𝗼𝗰𝗮𝘁𝗶𝗼𝗻𝘀
Our summer retreat 2026 wrapped up across 8 locations: Munich, Karlsruhe, Berlin, Hamburg, Cologne, Budapest, Oxford & Melbourne.
We discussed 75+ topics like the impact of AI on software development, agentic AI, security, code reviews and local AI models.
Our summer retreat 2026 wrapped up across 8 locations: Munich, Karlsruhe, Berlin, Hamburg, Cologne, Budapest, Oxford & Melbourne.
We discussed 75+ topics like the impact of AI on software development, agentic AI, security, code reviews and local AI models.
July 13, 2026 at 6:40 AM
𝗧𝗡𝗚 𝘀𝘂𝗺𝗺𝗲𝗿 𝗿𝗲𝘁𝗿𝗲𝗮𝘁 𝟮𝟬𝟮𝟲 𝗮𝘁 𝗲𝗶𝗴𝗵𝘁 𝗹𝗼𝗰𝗮𝘁𝗶𝗼𝗻𝘀
Our summer retreat 2026 wrapped up across 8 locations: Munich, Karlsruhe, Berlin, Hamburg, Cologne, Budapest, Oxford & Melbourne.
We discussed 75+ topics like the impact of AI on software development, agentic AI, security, code reviews and local AI models.
Our summer retreat 2026 wrapped up across 8 locations: Munich, Karlsruhe, Berlin, Hamburg, Cologne, Budapest, Oxford & Melbourne.
We discussed 75+ topics like the impact of AI on software development, agentic AI, security, code reviews and local AI models.
How #AI is supporting #fiber network expansion. Together with #OXG, we built an AI-powered application that helps validate civil engineering services using image and geospatial data. Read more 👉 www.tngtech.com/en/about-us/.... Watch the full talk 👉 youtu.be/gerVVq8zT6M?...
July 6, 2026 at 9:51 AM
How #AI is supporting #fiber network expansion. Together with #OXG, we built an AI-powered application that helps validate civil engineering services using image and geospatial data. Read more 👉 www.tngtech.com/en/about-us/.... Watch the full talk 👉 youtu.be/gerVVq8zT6M?...
At BTD26, our colleagues covered a variety of topics: leveraging #AI agents for product owner work and #codemigrations, the newest findings from our AI research team, various #robotics experiments, and more. You can find a summary of selected talks here 👉 www.bigtechday.com/en/big-techd... #tngbtd
July 2, 2026 at 9:54 AM
At BTD26, our colleagues covered a variety of topics: leveraging #AI agents for product owner work and #codemigrations, the newest findings from our AI research team, various #robotics experiments, and more. You can find a summary of selected talks here 👉 www.bigtechday.com/en/big-techd... #tngbtd
Experience four days of future at the Festival der Zukunft at Deutsches Museum Munich. 🤖 TNG will be present with showcases and talks in the field of #AI, #robotics, and #gaming. Meet our robots, try Playspawn & AI Photobooth. More info 👉 www.1e9.community/festival-der...
June 29, 2026 at 11:29 AM
Experience four days of future at the Festival der Zukunft at Deutsches Museum Munich. 🤖 TNG will be present with showcases and talks in the field of #AI, #robotics, and #gaming. Meet our robots, try Playspawn & AI Photobooth. More info 👉 www.1e9.community/festival-der...
Tired of folding laundry? Our #OpenArm robot can take over. With only 100 demonstrations collected via our #VR teleoperation setup we fine-tuned a pi0.5 robot foundation model for this task. The teleop and inference pipeline is implemented in ROS 2. Full video 👉 youtu.be/kCcms6c7hu4
June 26, 2026 at 6:43 AM
Tired of folding laundry? Our #OpenArm robot can take over. With only 100 demonstrations collected via our #VR teleoperation setup we fine-tuned a pi0.5 robot foundation model for this task. The teleop and inference pipeline is implemented in ROS 2. Full video 👉 youtu.be/kCcms6c7hu4