iiinova — where your digital self keeps evolving

Or join us to hear about our own R&D work at the Workshop on Conversational Search for Complex Information Needs. convsearch-complex-info-needs.github.io #ecir2026

010

At @zeta-alpha.bsky.social, we integrate cutting-edge IR research into VLM-based RAG, Deep Research & AI agents. Proud to sponsor #ECIR2026! Visit our booth (yes! We are hiring... www.zeta-alpha.com/about) #ecir2026

220

The real differentiator in enterprise AI? Reliable retrieval of evidence across messy unstructured content, fragmented systems, and strict access rules. Robust retrieval powers RAG & agentic AI workflows with grounded answers, traceable citations, and permission-aware outputs. #ecir2026

100

Zeta Alpha @zeta-alpha.bsky.social · 26/02/2026

Excited to connect w/ IR community at @ecir2026.eu in Delft this week! R&D on search, relevance & evaluation is the foundation any advanced AI deployment in knowledge-heavy domains. The future? Agentic search systems that plan, query, evaluate & iterate, built on proven IR fundamentals. #ecir2026

152

LLM Evals become the engine for self-improving systems that outperform hand-written prompts and multi-agent systems. @abcamara.com talks Friday at @irlab-amsterdam.bsky.social's SEA on how Deep Research agents evolve without human intervention. www.meetup.com/sea-search-e...

051

While many are still stuck in the discussion phase, we’re delivering a blueprint for European AI independence, run by strong local businesses. 6/6 Read the full Press Release: www.zeta-alpha.com/post/oediv-a...

020

The whole thing runs on Zeta Alpha’s software, giving you state-of-the-art Deep Research, RAG, and AI agents, and it is built to securely tap into the internal knowledge hidden in your company’s own documents, making sure every answer is grounded in facts you actually own. 5/6

Everything stays in OEDIV’s high-security centers, protected from the US CLOUD Act. This specific setup is "Compliance by Design" so even high-risk use cases can go live while staying strictly within GDPR and AI Act rules. 4/6

Together with OEDIV, we are launching a new sovereign AI infrastructure hosted entirely in Germany. It’s the best of both worlds: the steady foundation of a trusted local family business combined with the agility of a pioneering AI startup, turning advanced sovereign AI into a practical reality. 3/6

In 2026, European organizations are increasingly forced to rethink their strategic IT choices and data sovereignty, while at the same time pushing to accelerate their AI adoption, and facing heightened requirements around compliance and risk management. 2/6

It takes local champions that act to build digital sovereignty. That’s why Zeta Alpha is teaming up with OEDIV, part of the Dr. Oetker Group, to deliver Sovereign AI and to securely connect companies’ internal knowledge with locally run AI models. 🧵 1/6 Reach out via lnkd.in/dktJM9W to learn more.

352

Check out the release notes via the Zeta Alpha blog. www.zeta-alpha.com/post/zeta-al... Enjoy Discovery!

010

Our new release allows your teams to not only create and schedule, but also to share powerful enterprise-ready AI Agents that connect securely to your internal knowledge, public research, and the web. Zeta Alpha empowers your teams to achieve more together, and faster.

100

The great thing with AI agents is that they can do work in the background while you are busy doing something else, which is why we have built controls so that you are able to schedule them and they send you a notification when they are done.

100

Everyone is experimenting with AI Agents, but few are able to show real success yet. One of the main reasons is that AI Agents need a secure connection to your company’s internal and external knowledge sources, and that’s exactly where Zeta Alpha comes in.

100

Zeta Alpha @zeta-alpha.bsky.social · 24/11/2025

Zeta Alpha launches custom AI agents built for knowledge-intensive enterprise teams. Easily create, share, schedule and manage your own AI agents that get real work done.

143

This month’s Trends in AI webinar focused on something that's rapidly moving from hype to reality: AI for scientific discovery, especially in chemistry, pharma, and materials science. Read more about it here: www.zeta-alpha.com/post/trends-...

AI & Scientific Discovery - Trends in AI: November 2025

This month’s Trends in AI webinar focused on something that’s rapidly moving from hype to reality: AI for scientific discovery, especially in chemistry, pharmaceuticals, and materials science.

010

Zeta Alpha @zeta-alpha.bsky.social · 19/11/2025

We'll cover how AI agents are accelerating research today, how close LLMs are to exploring new scientific ideas, and how they can support researchers’ knowledge work. As always, we'll round it out with the latest news in AI R&D, open-source releases, and the most trending papers of the month!

000

Zeta Alpha @zeta-alpha.bsky.social · 19/11/2025

Join us tomorrow at 5PM CET / 8AM PT, for a special episode of our Trends in AI show, with a focus on AI for scientific discovery in chemistry, life sciences, and pharma. Sign up: lu.ma/trends-in-ai...

lu.ma

Trends in AI by Zeta Alpha - Chemistry & Scientific Discovery · Zoom · Luma

Join us for the Zeta Alpha Trends in AI webinar on Thursday, November 20th, at 8 AM PT / 5 PM CET. The event is in person in Amsterdam (LAB42, Science Park),…

100

Reposted by Zeta Alpha

Good advice. Don’t select AI vendors based on capital raised, news headlines or Gigawatts of compute, but on ability to partner and to deliver.

Reposted by Zeta Alpha

The most interesting takeaway from the keynote? 84% of IT leaders don’t track accuracy. How can you govern and improve something that you don’t measure.

Reposted by Zeta Alpha

Also cool: Context Engineering, the fastest hype cycle rise from Andrej Karpathy tweet to Gartner CIO strategy advisory in less than 5 months. (With due credit for Tobi Lutke)

Reposted by Zeta Alpha

Zeta Alpha @zeta-alpha.bsky.social · 29/10/2025

Reporting for @zeta-alpha.bsky.social from Gartner Xpo Barcelona this week, and rest assured: it’s all about AI agents #GartnerSYM

331

RAG in production is no longer "retrieve a few documents, paste into a prompt, and hope the answer is correct". The modern pattern is Agentic RAG: plan-execute loops, search across enterprise sources and the web, shared memory, tool use, and verification. ✍️ Read more: zeta-alpha.com/post/agentic...

Agentic RAG: From Overviews to Deep Research

020

Reposted by Zeta Alpha

Zeta Alpha @zeta-alpha.bsky.social · 13/10/2025

For enterprises and organizations deploying AI apps in production, AI Safety is no longer optional. And by AI Safety we don't mean some abstract "AI risk", but the practical policies, controls, and engineering that keep systems secure, compliant, and trustworthy. 👉 www.zeta-alpha.com/post/enterpr...

Enterprise AI Safety, Without the Hype - Trends in AI: October 2025

For enterprises and organizations deploying AI applications in production, AI Safety is no longer optional. And by AI Safety we don't mean some abstract "AI risk", but the practical policies, controls...

032

Zeta Alpha @zeta-alpha.bsky.social · 13/10/2025

Enterprise AI Safety, Without the Hype - Trends in AI: October 2025

032

Zeta Alpha @zeta-alpha.bsky.social · 06/10/2025

We will dive into: 🔒 Secure AI infrastructure & data sovereignty 🧱 Guardrails & moderation for LLM outputs 💥 Jailbreaks & prompt injections in the wild 📜 Regulation like SB 53 and ISO 42001 compliance 🧠 Information integrity & the “AI workslop” productivity killer

Zeta Alpha @zeta-alpha.bsky.social · 06/10/2025

Join us this Friday, October 10th at 8 AM PST / 5 PM CEST for a special episode of our “Trends in AI” show, with a focus on Enterprise AI Safety.

100

Zeta Alpha - AI Research Navigator

Check out the full EmbeddingGemma paper on Zeta Alpha, where you can easily explore more related work using the "Find similar" functionality. Until next time, enjoy discovery! search.zeta-alpha.com/documents/4b...

search.zeta-alpha.com

The Zeta Alpha Neural Discovery Platform is a smarter way to discover and organize knowledge in AI and beyond, using Generative LLMs and advanced Neural Search.

000

Performance ✦ SOTA on MTEB multilingual/English/code ✦ Minimal int8/int4 quantization degradation ✦ Graceful drop from 768d to 128d, still leading its class ✦ Cross-lingual: beats many larger open & API models

100

Tricks ✦ "Spread-out" regularizer for quantization robustness ✦ Model souping across diverse data mixtures ✦ Quantization-aware training & MRL during finetuning

100

Training / Architecture ✦ Turn Gemma 3 into an encoder-decoder LLM using the UL2 objective (like GemmaT5) ✦ Keep the encoder, pre-finetune it on a large unsupervised dataset, then finetune on retrieval datasets with hard negatives ✦ Embedding distillation from Gemini Embedding

100

What is EmbeddingGemma? ✦ 300M parameter text embedding model ✦ derived from Gemma 3 ✦ SOTA on multilingual, English & code retrieval tasks ✦ matching performance of models 2x its size ✦ supports dimensionality reduction ✦ comes with quantization-aware checkpoints

100

Zeta Alpha @zeta-alpha.bsky.social · 18/09/2025

DeepMind just dropped the technical report behind their SOTA open-weights embedding model, packed with training details, including ablations on modeling choices like encoder-decoder initialization, mean pooling, and model souping. Let's take a look! 📚

120

Shipping a GenAI solution in production with a clear path to ROI is hard. In our latest post, we outline key challenges in Enterprise RAG, go beyond the basics with agents, and explore how automatic optimization can adapt these systems to domain-specific needs. www.zeta-alpha.com/post/why-gen...