Oleg O.

I am proud to announce Amazon is planning to invest $10 billion in North Carolina to expand our infrastructure to support AI and cloud computing technologies. The investment in the state is also expected to create at least 500 new high-skilled jobs and support thousands of other jobs in the Amazon Web Services supply chain. When Amazon data centers become a part of local communities, they serve as catalysts for opportunity, sparking job creation, increasing household incomes, and driving economic growth. If you’re curious to learn more: https://lnkd.in/gYdXW2Yw

705

4 Comments

Exciting news from Oracle Labs! We’re pleased to announce the open source release of the Open Agent Specification, a major step forward in building an open and interoperable AI agent ecosystem. The Open Agent Specification provides a standardized, declarative framework for defining AI agents and workflows, and is accompanied by robust tooling and a reference implementation. This initiative aims to foster collaboration, innovation, and reusability across the AI and developer communities. Huge thanks to our partners in Database, OCI, and Global Industry Units for their contributions and collaboration! Learn more and get involved in the project here: https://lnkd.in/gjKpD5VC #AI #OpenSource #AIAgents #OracleLabs #OpenAgentSpecification

261

6 Comments

Thrilled to have been part of an insightful panel at the Texas AI Summit with Mustafa Ali, Kayla Ball, and Mehrab Momin, and guided brilliantly by Nada Ahmed. We explored topics ranging from AI adoption and cybersecurity to the evolving AI infrastructure landscape. It was a dynamic discussion with thoughtful audience questions throughout. I shared my perspective on sovereign AI, which sparked meaningful follow-up conversations with attendees who recognized it as an increasingly urgent need. For enterprises operating in sensitive domains such as defense, healthcare, finance, and advanced R&D, compliance through virtual firewalls is no longer sufficient. Achieving true sovereignty requires secure, physically isolated compute capacity that ensures complete control over data and infrastructure. I also emphasized that cybersecurity must extend across both software AND hardware layers, a crucial step as enterprises work to protect their most sensitive information. These conversations reaffirm the growing importance of aligning innovation with trust, control, and resilience. Curious to hear how others are approaching data sovereignty and AI security in their organizations.

114

3 Comments

We are expanding our partnership to bring Google’s leading Gemini models to Oracle Cloud Infrastructure (OCI) Generative AI service via Vertex AI. This is a huge step in making our advanced AI more accessible, allowing businesses to build powerful AI agents and streamline complex tasks right within their Oracle environment. Looking forward, we will also collaborate with Oracle to make Gemini models available within their Fusion Cloud Applications, which will enhance workflows for teams in finance, HR, supply chain, and more. It’s fantastic to see this level of collaboration come to life, and I'm genuinely excited to see the innovative solutions our joint customers will create. A great day to be part of the team at Google! #GoogleCloud #Oracle #Partnership #GenerativeAI #AI #TeamGoogle https://google.smh.re/55WQ

1,734

54 Comments

NEW RELEASE from Packt Publishing >> "LLMs in Enterprise: Design strategies, patterns, and best practices for large language model development" See it at https://amzn.to/4pJK98J 𝓦𝓱𝓪𝓽 𝓨𝓸𝓾 𝓦𝓲𝓵𝓵 𝓛𝓮𝓪𝓻𝓷: 🟠Apply design patterns to integrate LLMs into enterprise applications for efficiency and scalability 🔵Overcome common challenges in scaling and deploying LLMs 🔴Use fine-tuning techniques and RAG approaches to enhance LLM efficiency 🟠Stay ahead of the curve with insights into emerging trends and advancements, including multimodality 🔵Optimize LLM performance through customized contextual models, advanced inferencing engines, and evaluation patterns 🔴Ensure fairness, transparency, and accountability in AI applications

19

3 Comments

Is your enterprise infrastructure truly prepared for the era of AI agents? Gartner’s AI Infrastructure Report (2025) reveals a telling insight — 86% of enterprises deploying AI agents acknowledge their current infrastructure isn’t ready. The core issue isn’t just technology; it’s alignment. Traditional CI/CD pipelines were designed for deterministic software, not for autonomous systems that evolve through prompts, reasoning loops, and live API interactions. As organizations scale their AI initiatives, this gap becomes increasingly evident. Agentic systems require new forms of validation and governance, including: • Continuous prompt and reasoning loop testing • Red-team assessments for safety and reliability • Shadow deployments, canary rollouts, and automated rollback strategies Enterprises that succeed — like those leveraging Microsoft’s AutoGen framework (adopted by 40% of Fortune 100 firms) — approach agent deployment as an ongoing validation cycle, not a one-time release. The shift from static to autonomous systems is redefining how infrastructure, DevOps, and AI governance intersect. 💡 How prepared is your organization to operationalize this shift — both technically and strategically? Cc: Bhavishya Pandit

149

9 Comments

The MCP Architecture Stack That Enables Autonomous SoR Agents Building enterprise AI agents is hard. Semantic accuracy over messy data. Business context across tribal knowledge. Grounding in actual system state rather than hallucinated responses. Solve all that, and there's another problem: they still can't act. Detection: ✓ (hard) Diagnosis: ✓ (hard) Recommendations: ✓ (hard) Execution: ✗ (harder) The last mile makes everything more complicated. Each agent needs custom integration code for every system it touches. CRMs, ERPs, billing, inventory - all with different APIs, authentication, error handling. N × M complexity that explodes as agent count grows. Model Context Protocol addresses this with standardized integration architecture. Instead of point-to-point chaos, MCP creates a protocol layer. Agents expose capabilities. Systems expose actions. Complexity shifts from N × M to N + M. Collections Agents escalate high-risk accounts automatically across systems. Inventory Agents trigger replenishment orders without human relay. Sales Ops Agents coordinate pricing adjustments in real-time. The progression: Read operations → Write with approval → Autonomous execution The value: 100x speed, consistent process every time, platform effects that compound The safety: Rate limits, approval thresholds, semantic constraints, rollback capabilities This is the architectural shift from AI advisors to AI operators. From agentic systems to autonomous enterprises. Here is the new post on my medium: https://lnkd.in/eYg4G4H2

20

We built @Parasail around four principles that are critical for AI builders today. Context: most infrastructure providers compete on scale. Their business model is built on enormous contracts, racks of GPUs, and long deployment timelines. That works for enterprises ready to spend hundreds of millions upfront. But most AI teams need something else: flexibility, speed, and reliability at a scale that makes sense for them. Here’s how we designed Parasail to serve those needs: (1) Contract flexibility. Not every team can commit to a $100M contract. Many need to spend $500K or $1M a year and grow from there. We support those customers without forcing them into oversized deals. (2) The full inference stack. Hardware alone isn’t enough. Running production inference requires deployment pipelines, scaling logic, monitoring, and optimization. We deliver that as part of the stack, so teams don’t need to build DevOps from scratch. (3) Speed to production. Most providers take months to get customers live. We move from first conversation to production traffic in days. That speed is essential when AI teams are racing the clock. (4) White-glove support. Instead of a GPU catalog where customers pick blindly, we work side by side with them. Our engineers analyze the model, recommend the right configuration, and tune it for performance. Our goal is to give AI teams the flexibility, reliability, and speed they need so they can focus on building products, not battling infrastructure. We’re here to make sure inference is never the bottleneck to innovation.

36

1 Comment

Greengage Database (based on Greenplum®, open source alternative)(https://lnkd.in/ehaCaYHt) is now available right inside DBeaver. Starting with the latest release on October 6, you can connect to and work with it directly from the DBeaver interface - just like you do with PostgreSQL, Greenplum, or any other database.

26

Light vs. heavy agents miss the real axis: durability vs. disposability. I was thinking about this while reading Hamza's latest post. He makes a compelling distinction between Light Agents (chat wrappers) and Heavy Agents (orchestrated workflows). He's right that enterprise reliability needs orchestration, sandboxing, and state management. But the Light vs Heavy axis describes complexity. It misses the more critical question: Durability vs Disposability. Traditional batch workflows were built to last. These are capital investments measured in years or decades. Agents are different. Reviewing my own "Heavy" agent architectures from 2024, I've found that robustness is often an expensive trap. Many "Heavy" agents built 18 months ago were sophisticated band-aids. They used complex chaining and error handling to compensate for model stupidity. Today, a single call to a frontier model often outperforms those brittle, multi-step chains. If you spent 6 months productionizing a workflow that the next model update solves natively, you didn't build an asset. You built technical debt. When to Go Heavy? Build Heavy only when: - The process is high volume and low variance - Compliance requires an audit trail - The cost of failure exceeds the cost of engineering Everything else? Stay light. Traditional automation is optimized for repeatability. Agentic automation optimizes for leverage in the face of uncertainty. Models improve faster than architectures age. That inverts the traditional calculus of software engineering. In 2010, you built once and ran for years. In 2026, you might build something that's obsolete in six months because the model got smarter. That doesn't mean you built wrong. It means you built for the reality that existed, and you need to be ready to rebuild when reality shifts. How are you calculating ROI on your agents? Are you factoring in the "Model Obsolescence Rate"? Are you building for durability when you should be building for velocity? Most importantly: are you optimizing for the system you have today, or the system you'll have in Q3? Because if you're building "Heavy" for a problem that Claude 4 or GPT-5 solves natively, you're not future-proofing. You're past-anchoring. Follow Rod Rivera for more on agent engineering, AI operations, and building in a moving landscape.

26

3 Comments

As a launch partner for OpenAI gpt-oss models, Databricks is bringing their new gpt-oss-120b and gpt-oss-20b models directly to the Databricks platform. Securely build custom AI agents with state-of-the-art open models, all with the full governance and flexibility of Databricks. Fine-tune on your private data and deploy for any use case, from summarization to real-time agents. Databricks' post - https://lnkd.in/eP5TkQtg OpenAI's post - https://lnkd.in/etxskikp #databricks #openai #gpt #ai #gpt-oss #agentbricks

46

Excited to share Informatica’s latest press release and highlight three Generative AI boosters: BETTER, SAFER, FASTER. Full press release: https://lnkd.in/g25WdCDz SUMMARY AND QUICK QUOTES: BETTER GenAI is only as good as the data it learns from. Informatica now makes it easier to clean up messy data, fill in missing details, and explain how information is grouped or matched. This means your AI can provide more accurate, reliable, and understandable answers—especially when using techniques like RAG (retrieval-augmented generation) or connecting to company-specific knowledge. “CLAIRE® Match Analysis and Explainability: Provides comprehensive match analysis with explainability for transparency into how records are matched" “Enrichment and Validation Orchestrator: MDM validation and enrichment… across Informatica services, third-party data sources, and large language models (LLMs).” “CDGC + MDM integration… improves auditability of master data assets, reduces data silos and gets data ready for AI and analytics.” SAFER As companies adopt AI, it's important to keep track of what data and models are being used—and who approved them. With these new features, teams can monitor AI models like any other business tool. You’ll be able to track where the data comes from, who made changes, and ensure that the information feeding into AI tools is clean and safe. “AI Governance Inventory & Workflows… tracks models, datasets, and approvals, ensuring responsible and compliant deployment” “AI-Powered Lineage Discovery… maps data flows from source systems to applications and AI models” “Data Quality Rules as API… customers can assess, clean and enrich their data in real-time.” FASTER To move quickly with AI, developers need fast access to the right data—no matter where it lives or what format it's in. Informatica now helps developers discover usable data faster, connect to tools like NVIDIA and Databricks, and reuse templates to get started quickly. Whether building chatbots, image generators, or internal copilots, teams can skip the heavy lifting and focus on solving real problems. “Recipe Marketplace: Users can now publish and share reusable workflow templates ‘recipes’ ranging from GenAI use cases like NVIDIA image generation and Slack AI Agent RAG…” “Expanded connectivity… includes catalog integration for Apache Iceberg, SAP HANA CDC, and Microsoft Fabric Delta Tables.” “Metadata Scanner… catalogs MDM assets, maps data lineage, adds business context and enforces compliance requirements.” COMING UP In our upcoming Tech Talk, we’ll provide a one-hour live walkthrough of how Informatica's IDMC platform powers trusted, enterprise-ready Generative AI. Join us to see how IDMC makes GenAI BETTER, SAFER, FASTER. #GenAI #BetterSaferFaster #Informatica #AIReadyData #ResponsibleAI #DataTrust #IDMC

17