Dania Kodeih

A recent Business Insider perspective highlights how AI is starting to blur the line between digital and physical systems. What used to be a software-led advantage is now expanding into infrastructure, energy, and real-world operations. That resonates with what many organizations are experiencing. Data and analytics no longer sit in isolation. They are tied directly to physical assets, supply chains, and energy systems that enable scale. As AI makes digital output easier to produce, the differentiator is how effectively organizations connect data to real-world execution. From my perspective, the next phase of advantage comes from integrating analytics with operational reality. The closer data gets to execution, the more impact it delivers.

9

I had the honor of co‑hosting the keynote at CODE2026, ExxonMobil’s developer conference in Houston, opening with this retro‑style slide. It was there to anchor the point that AI models don't “𝘸𝘢𝘯𝘵” anything and don’t originate direction — the system that we develop does. LLMs remain operationally opaque: probabilistic models without deterministic traceability. We moved swiftly from single‑shot prompts into RAG because it was the only scalable way to anchor generative systems in enterprise truth. And retrieval didn’t stay a bolt‑on; it became infrastructure, the substrate for policy grounding, decision support, tool execution, memory and context continuity. Assertion‑based testing still applies everywhere the system behaves deterministically, but cannot validate probabilistic model behavior, so evaluation‑driven development emerged to supervise the parts of the system where exact‑match assertions break down. Once evals become the feedback loop, the work shifts from prompt‑tuning to system‑level behavior design. That evolution pushed architectures toward agentic patterns. Agents plan, reason, call tools and act across multiple steps, which means they accumulate uncertainty and propagate error unless they remain supervised, observable, interruptible. Multi‑agent systems reintroduced distributed‑systems problems: concurrency, messaging, state, failure modes... except the nodes can now be stochastic. That’s the real shift. Not prompt engineering, but 𝘀𝘆𝘀𝘁𝗲𝗺𝘀 𝗲𝗻𝗴𝗶𝗻𝗲𝗲𝗿𝗶𝗻𝗴 𝘄𝗶𝘁𝗵 𝗽𝗿𝗼𝗯𝗮𝗯𝗶𝗹𝗶𝘀𝘁𝗶𝗰 𝗰𝗼𝗺𝗽𝗼𝗻𝗲𝗻𝘁𝘀. And exactly because these systems aren’t “𝘢𝘭𝘪𝘷𝘦”, any coordination, supervision, uncertainty‑resolution or failure‑handling they exhibit comes from the structures we design — not from intrinsic intent in the models. As autonomy increases, the engineering surface area increases. The narrative that AI reduces the need for developers misunderstands the architectural reality.

218

14 Comments

One of the challenges of modern Cloud Compute is the complex interdependence on "whole of stack" Cloud vendors. AWS knows how to operate data centers and hyperscale. ...but for its Cloud customers, far too often, the behind the scenes "digital duct tape" gets exposed via outages. and in nascent industries - such as Quantum, or Crypto, these cracks in the infrastructure can mean outsized negative impacts. https://lnkd.in/eVdenexN

5

With increasing energy demands and tighter capital budgets, companies are turning to smarter, more efficient ways to optimize their existing assets. Check out a new article written by Baker Hughes Director of Digital Operations, Omar Iskandar, to explore the vast capabilities of Leucipa's Live Simulation Engine. At the heart of this transformation is the digital oil field—where real-time data and advanced engineering models come together to unlock new levels of performance. Read the full article here: https://lnkd.in/du7fniaN #DigitalTransformation #OilandGas #OilandGasIndustry #WeAreBakerughes

3

🚀 Driving Organizational Transformation through Data & AI 💡 Oil & Gas | Banking | Healthcare | Government & Public Sector Data is the new oil. AI is the engine fueling innovation, efficiency, and citizen-centric services. From offshore rigs and operating rooms to government offices and banking systems, organizations are leveraging data and AI to rethink how they operate, serve, and compete in a rapidly evolving landscape. 🔍 Key Questions Every Leader Should Ask: 1. How can we streamline processes using data and AI? 2. Are we truly ready for change — culturally and technologically? 3. What role should AI play in augmenting human decision-making? 4. How do we reduce vulnerability to competition and disruption? 5. How can AI help reduce cash flow volatility and improve long-term resilience? ✅ Oil & Gas: Predictive maintenance, drilling optimization, emissions monitoring. ✅ Banking: Smart credit, fraud prevention, hyper-personalized experiences. ✅ Healthcare: AI-assisted diagnostics, efficient patient care, smart scheduling. ✅ Government & Public Sector: Digital citizen services, AI-powered policy insights, optimized resource allocation. 📊 The public and private sectors alike must embrace data-driven thinking to unlock value, drive transparency, and serve with agility. The future belongs to those who innovate with purpose. Is your organization ready? #AI #DigitalTransformation #PublicSector #SmartGovernment #OilAndGas #Banking #HealthcareInnovation #DataDriven #ArtificialIntelligence #Leadership #FutureOfWork #AIforGood #isb

3

In the Era of industrial #AI, Cybersecurity Is Not Optional , It's Foundational! As industries embrace AI to unlock new levels of efficiency and intelligence, securing operational technology (OT) environments becomes more critical than ever. The convergence of IT and OT introduces both opportunity and risk especially when AI systems begin accessing sensitive OT data. There is no autonomy without cybersecurity. AI-driven operations can only succeed when built on a solid foundation of cyber resilience. From predictive maintenance to remote operations and autonomous, every layer of digital innovation depends on trust and trust starts with security. At Kent, our industrial cybersecurity team brings domain-certified expertise to ensure your operations evolve safely and securely. We help you: ✅ Secure data access across IT/OT environments ✅ Assess vulnerabilities introduced by AI systems ✅ Implement governance frameworks tailored to industrial environments ✅ Achieve compliance without compromising performance Let’s talk about securing your path to autonomous operations. #CyberSecurity #IndustrialCyberSecurity #AIinIndustry #OperationalTechnology #OTSecurity #ITOTConvergence #AutonomousOperations #DigitalTrust #Kent Manish Thakur Eiman Alhashmi

40

Lean into AI, to code, even for architects. We are working on an architecture for one of our enterprise integration platforms. The strategy is sound, but I want to personally test certain scenarios without slowing down the team.  By leveraging the latest Frontier and Premium LLMs, I was able to iterate through the following: build the core idea, create cloud infrastructure mocks and service mocks, run tests, analyze results, and continue the feedback loop.  Ultimately, this equips me with the insights to step into the "engine room" and collaborate with the delivery team.  Lean into AI.

31

4 Comments

GhostGPT is an uncensored generative AI chatbot designed explicitly for cybercriminal activities. Unlike mainstream AI models that incorporate ethical guidelines and content filters, GhostGPT operates without such restrictions, enabling users to generate malicious content with ease. https://bit.ly/3EcmOtm

5

I strongly believe that SOC doesn’t need more alerts. It needs better decisions. Burnout isn’t caused by volume alone. It’s caused by low-confidence signals. High-performing SOCs focus on: - Fewer detections - Clear response paths - Measurable outcomes Automation without clarity = faster chaos. #SOC #ThreatDetection #SecurityOperations #CyberDefense #CyberSecurity

2