SOTI

Tomorrow we're sitting down with Peter Turpin, CEO of Surveil, to talk about a role most enterprises don't have yet — but probably should. Chief Cloud Intelligence Officer. 25 minutes. No fluff. Just an honest conversation about why cloud financial management has outgrown the org structures it currently lives in. If you're in IT, finance, or engineering and you've ever been caught off guard by a cloud bill — this is worth 25 minutes of your Tuesday. Register link in the comments. #FinOps #CloudCost #EnterpriseIT #CCIO

What happens when you hand 100+ engineers a map of a 15-million-line codebase? Chaos? No. Clarity. First — a genuine thank you. None of this would have happened without the trust and backing of Christophe Roux , Yoons Alumvilayil Srajudeen and Manu Xavier . You gave me the mandate, the resources, and the runway to go build something the team hadn't seen before. That means more than I can put into words. Thank you. Last week I ran a Tribe-wide knowledge-sharing session — 100+ attendees: developers, senior architects, QA Automation engineers, and Product Owners — all gathered around one question: "Can we actually understand a 15 million+ Lines of Code monolith well enough to let an AI reason over it intelligently?" The answer we built: a Knowledge Graph of the entire Monolithic codebase. Not a search index. Not a vector store. A proper, structured, traversable graph — loaded into Neo4j, hosted on a VM, queryable in real-time with Cypher. The numbers (verified from our pipeline benchmark): Over 2,031,000 nodes — every class, method, interface, EJB, DB table, SQL procedure, config entry Over 3,276,000 edges — every call, dependency, inheritance, read/write, JNDI binding, schema reference Parsed from 29,000+ Java files, 9,900+ SQL files, and 300+ XML configs — in under 10 minutes Why does this matter more than a fancy diagram? Here's the part people don't talk about enough: token budget. When you throw a 15M LOC codebase at an LLM agent without context management, you burn tokens fast — and you still get hallucinations, because the model doesn't really know what's connected to what. Knowledge graphs change that equation. Instead of feeding raw source code, you query the graph for exactly the structural context that's relevant — and pass that to the agent. Based on patterns in this space, this approach can reduce agentic token consumption by 35–60% while improving reasoning accuracy, because the model gets precise, relationship-rich context instead of a firehose of text. Think of it as giving the AI a GPS instead of a 15-million-page atlas. The tech stack: Neo4j · Docker · LangGraph · tree-sitter · Python · sqlparse · lxml I can't share proprietary details of the codebase or system — that goes without saying. But the learning I can share is this: Knowledge graphs aren't a research project. They're infrastructure. And when you build them right, they become the single most useful primitive for any AI agent that needs to reason over a complex system. We're just getting started. To the 100+ people who showed up, asked sharp questions, and pushed back constructively — you made that session worth running. Thank you. #KnowledgeGraph #Neo4j #LangGraph #AIEngineering #GenAI #SoftwareArchitecture #Agentic #Docker #GraphDatabase #EnterpriseAI #IBSGroup #IBS

Hiring: Sales Enablement Product Manager Intern (12-Month Co-op) at SOTI Looking for someone who: ✨ Wants to learn Product Management ✨ Has strong AI prompt engineering skills ✨ Knows basic C# / development fundamentals ✨ Loves building, experimenting, and solving problems You’ll work on Sales Enablement Products, AI-powered solutions, and real-world product challenges with a global team. If AI, product ideas, workflows, and innovation genuinely excite you — we should talk 👀 📍 Mississauga, ON Apply here: https://lnkd.in/exEsmUuM #Hiring #ProductManagement #AI #Internship #Coop #SOTI #Salesenablement

Farewell Melbourne! As we prepare for our move to Qld in the next few days, it is with mixed emotions I farewell my home town of almost 60 years. I know we will be regular visitors in the future, but as we leave, I have to recognise the wretched state of our once great city and state and it saddens me as we make the move. I implore each and every Victorian to ask themselves how they can make things better in the future. Volunteer at your kids sport, offer to help someone who needs assistance, stand up against poor behaviour and most of all, don't accept the dishonesty and deception we have been faced with in Spring Street. Victoria deserves better in the future than what we have seen over the last decade or so. I will always love Melbourne, the Sport, the Theatre, the Coffee, the G and most of all the people. This is not goodbye, but a farewell to my home city, something I really never thought possible only a decade or so ago, Cheers, Michael

🚀 Java Interview Revision Notes — JVM, Garbage Collection & Inner Classes Saving these important Java concepts/questions for future revision 📌 These topics are frequently asked in Java backend interviews and are super important for understanding JVM internals, memory management, and object-oriented design. 🔥 JVM & Garbage Collection Questions ✅ Strong Reference vs Weak Reference When should we use Strong Reference? When should we use Weak Reference? Advantages of each? Why do we need different types of references in Java? Difference between SoftReference, WeakReference, and PhantomReference? How WeakHashMap works internally? ✅ Heap Memory Structure What are Eden, Survivor spaces (S0, S1), Young Generation, Old Generation? Why are there so many spaces in Heap memory? How objects move from Young Gen to Old Gen? What happens during Minor GC and Major GC? Why is object aging important? ✅ Garbage Collection What is Mark and Sweep algorithm? What is Mark-Sweep-Compact? Why compaction is needed? What are other Garbage Collection algorithms? Why is Garbage Collection expensive? What causes Stop-The-World pause? How does GC improve memory management automatically? ✅ Types/Versions of Garbage Collectors Serial GC Parallel GC CMS GC G1 GC ZGC Shenandoah GC Questions: Which GC is default in modern Java? Difference between G1GC and ZGC? Which GC is best for low latency applications? Why CMS was removed? ✅ Metaspace Why do we need Metaspace? Difference between Metaspace and PermGen? What data is stored in Metaspace? Why was PermGen removed? Can Metaspace cause OutOfMemoryError? ✅ Class Variables What are class/static variables? Difference between instance variables and class variables? Where are static variables stored in memory? Why static variables are shared across objects? 💡 Consistency in revising core Java concepts builds strong backend fundamentals. #Java #JVM #GarbageCollection #JavaDeveloper #BackendDeveloper #CoreJava #JavaInterview #SoftwareEngineering #Programming #Developers #JavaInternals #CodingInterview #SystemDesign #JavaLearning #TechInterview #OOP #MemoryManagement #JavaBackend #RevisionNotes

Is AI becoming the healthcare product itself? Several recent pharma AI announcements focused on speed: faster drug discovery, faster clinical trials, faster operations, faster enterprise workflows. Novo partnered with OpenAI to accelerate the business. Lilly partnered with NVIDIA to accelerate molecule generation and R&D. But Roche just made a very different bet with its proposed $1.05B acquisition of PathAI , looking to embed AI into the clinical product layer itself: pathology, biomarker analysis, companion diagnostics, and clinical decision support. Operational AI can often be replicated. Deeply integrated clinical workflow infrastructure is much harder to replace. That is why Roche’s move stands out in the growing wave of AI partnership headlines. Roche is building the intelligence layer between diagnosis and therapy. A few others are making similar moves. Siemens Healthineers is embedding AI into imaging workflows. GE HealthCare is integrating AI into radiology platforms. Tempus built AI directly into oncology decision support and genomic matching. These companies are not just using AI to run faster. They are using AI to redefine the product. #healthcareAI #pharmaAI

🧠 Four years ago, around this time, I felt completely stuck. No clear direction. No confidence in my next step. Just movement without meaning. ✨ Today, things aren’t perfect, but they’re clearer. 🌱 What changed? ➡️ I stopped waiting for clarity and started taking action ➡️ I invested time in learning, not just working ➡️ I started putting myself out there (even when it felt uncomfortable) 💡 Clarity doesn’t come before action. It comes because of it. 🚀 If you’re feeling stuck right now - you’re probably closer than you think. Keep going. Keep showing up. Your future self is watching. 💛 #GrowthMindset #CareerGrowth #PersonalDevelopment #KeepGoing #ProgressNotPerfection

Java Intro Basics — But Think “Why?” 🤔 Most of us learn Java by memorizing definitions. But real understanding starts when you begin to question them. Here are some basic concepts — but instead of answers, ask yourself why: 🔹 Java is called platform-independent (WORA) 👉 Why, when Windows, Mac, and Linux all have different JDKs and JVMs? 🔹 JVM is an abstract machine 👉 Why isn’t it physical? What problem does that solve? 🔹 JVM is platform-dependent 👉 Then how does Java remain platform-independent at the same time? 🔹 Java Program → Bytecode → JVM → Machine Code 👉 Why introduce bytecode at all? Why not compile directly to machine code? 🔹 JVM has a JIT compiler 👉 Why compile at runtime? Isn’t compilation already done by javac? 🔹 JRE vs JDK 👉 Why can we run a program with JRE but not develop one? 🔹 JDK = JRE + tools 👉 Why separate runtime and development environments? 🔹 main(String[] args) 👉 Why does JVM pass arguments as an array of Strings? 🔹 File name = public class name 👉 Why does Java enforce this rule? 🔹 One public class per file 👉 Why restrict it? What would break otherwise? 🔹 Java Editions (JSE, JEE, JME) 👉 Why different editions instead of one unified platform? 💡 The difference between a beginner and a strong developer is simple: Beginners memorize. Developers question. Engineers understand. Start asking why — that’s where real learning begins. #Java #Programming #Coding #Developers #interview #Learning #TechCareers

Hey #connections 👋 "𝗧𝗵𝗲 𝗺𝗼𝗿𝗲 𝗜 𝗽𝗿𝗮𝗰𝘁𝗶𝗰𝗲, 𝘁𝗵𝗲 𝗹𝘂𝗰𝗸𝗶𝗲𝗿 𝗜 𝗴𝗲𝘁" It’s been a while since I’ve been using SQL, but the opportunities coming my way often revolve around joins, partitions, CTEs, and more. Even in the data world, SQL is foundational—it supports Data Engineers, Data Analysts, BI Engineers, and all data enthusiasts. Being future-focused is always a good idea, so I decided to sharpen my SQL skills. Currently, I’m working in Microsoft Fabric, sleeping with PySpark, notebooks, and pipelines, and SQL is sitting in another room😄 Since I don’t have much time to research everything in depth, I picked up a course on LinkedIn and found a real gem💎 It’s an awesome course that covers many advanced topics, and the instructor, Dan Sullivan, explains everything perfectly. Thank you, Dan Sullivan , for such a sharp and insightful learning experience. I highly recommend this to my friends—if you’re at an intermediate level in SQL, go for it. It will definitely take you to the next level🚀 . . . . Course Link: https://lnkd.in/gi_gpkFq #AdvancedSQL #SQLLearning #DataSkills #Upskilling #LearnSQL #CareerGrowth #ContinuousLearning #DataEnthusiast #DataAnalytics #DataEngineering