Sushil Suri

Why “moving fast” is slowing your engineering programs down –my personal view; opinions are my own. Most delays in engineering programs are not caused by a lack of agility. They’re caused by decisions that were never verified early enough. When things slip, the V-Model is often blamed. “Too slow.” “Too heavy.” “Not iterative.” That diagnosis is convenient. And usually wrong. The real issue is not the V-Model. It’s the confusion between speed and validated progress. What I’ve seen repeatedly across large programs: • Teams optimize for output instead of evidence • Assumptions are made implicit to “go faster” • Interfaces are agreed late, under pressure • Verification is postponed until integration exposes reality The result looks like velocity at first. Then integration hits. And cost, risk, and headcount grow exponentially. Used properly, the V-Model is not a delivery model. It’s a governance mechanism. • Agile defines cadence • The V enforces alignment • Iteration happens inside small, focused loops • Verification moves risk forward in time—when it’s still cheap Programs that scale don’t rely on heroics at the end. They reduce uncertainty early and deliberately. Velocity without early verification is not agility. It’s financial risk disguised as progress. Where have you seen “speed” pay off—and where did it quietly turn into rework, delay, or recall? #engineeringleadership #systemsengineering #vmodel #softwarestrategy #deliveryexcellence

18

7 Comments

AI will make your teams faster. It will not make your structure stronger. Most SDV programs measure velocity: release cadence, feature throughput, OTA frequency. Very few measure absorption capacity. But every engineering system runs on two rates: – the rate of change generation – the rate of change absorption If generation exceeds absorption, instability accumulates. Not because of talent. Not because of tooling. Because structure cannot absorb propagation. When change moves faster than structure, cost compounds. In my latest Substack article, I explore why absorption capacity — not velocity — determines the economics of software-centred vehicles. If your cost-of-change curve is steepening, the constraint may not be where you think it is. Full article below. #SoftwareCentredVehicle #SDV #AutomotiveLeadership #PlatformStrategy #OTA

8

1 Comment

Google has introduced a transformative new AI experiment aimed at redefining how we learn. Their platform, “Learn Your Way,” has the potential to turn traditional textbooks into personalized, adaptive learning experiences – tailored to each individual’s pace, interests, and style. https://lnkd.in/gefMXdKK As AI continues to reshape education, innovations like these bring us closer to a future where learning is more engaging, inclusive, and truly learner-centric. Excited to see how this will empower students, professionals, and lifelong learners across the globe. #AI #EdTech #PersonalizedLearning #Innovation #FutureOfEducation

10

The day an engineering leader discovered she was “disposable.” —my personal view; opinions are my own. It was a Tuesday morning when my phone rang. Not a scheduled call. Not a casual check-in. A “Can you talk now? It’s urgent.” kind of message. When I answered, her voice was steady — but I could feel the weight behind it. She had just learned she was no longer needed in the company she had helped build. And she wasn’t a junior engineer. She was an engineering leader — the person people run to when everything is falling apart. Her pattern was always the same: show up, deliver, fix what’s broken, build teams, create stability. She didn’t just deliver results — she built the systems that made results possible. She even had a second-in-line ready, because responsible leaders think ahead. And still… with one leadership change, she became “disposable.” This was right after she landed the biggest customer acquisition of the year — the one everyone else had been chasing. The new leader didn’t ask about context. He didn’t ask about the architecture she corrected, the team she developed, or the crises she carried quietly so others could move forward. He looked only at the numbers printed on paper — and paper never tells the real story. She wasn’t dismissed for lack of talent. She was dismissed because someone didn’t care to see her impact. When she finished explaining, I paused. Not because she had no options — but because betrayal hits differently when you’re the one usually holding everyone else together. This is what I told her: “Take the hit without losing your center. Don’t argue for a seat that’s already been taken. Use your network. Show the market who you are again. The right doors always open when you move with clarity.” She followed the plan. And soon enough, the offers started coming. Not one. Not two. Three solid offers — all arriving before she even considered resigning. That changed everything. So when she finally walked into HR and handed in her resignation letter, she did it calm, confident, and fully in control of her next step. And then the twist: The same company that treated her as disposable suddenly tried to keep her. A counter-offer. A “let’s talk.” A late realization of her value. But it was already too late. She didn’t leave out of anger — she left out of clarity. The lesson? No title protects you from poor leadership. But clarity protects you from staying where you’re no longer seen. Know your worth. Watch the signs. And build relationships long before you need them. If this resonates and you want to go deeper, send me a DM. I’m opening a few mentorship slots for January. #mentorship #careerdevelopment #leadership #engineeringleadership #growth #knowyourworth

14

𝗜𝗳 𝘆𝗼𝘂𝗿 𝘁𝗼𝗼𝗹𝘀 𝗰𝗮𝗻’𝘁 𝗯𝗲 𝘁𝗿𝘂𝘀𝘁𝗲𝗱, 𝘆𝗼𝘂𝗿 𝘀𝘆𝘀𝘁𝗲𝗺 𝗰𝗮𝗻’𝘁 𝗯𝗲 𝘀𝗮𝗳𝗲. AI is changing how we build safety-critical systems — but most development pipelines now rely on tools that were never designed with functional safety in mind. ➤ Data labeling tools ➤ Model training frameworks ➤ Simulation or perception validation platforms 𝘜𝘯𝘥𝘦𝘳 𝘐𝘚𝘖 26262, 𝘵𝘩𝘦𝘴𝘦 𝘵𝘰𝘰𝘭𝘴 𝘮𝘶𝘴𝘵 𝘣𝘦 𝘲𝘶𝘢𝘭𝘪𝘧𝘪𝘦𝘥 𝘪𝘧 𝘵𝘩𝘦𝘪𝘳 𝘧𝘢𝘪𝘭𝘶𝘳𝘦 𝘤𝘰𝘶𝘭𝘥 𝘪𝘯𝘵𝘳𝘰𝘥𝘶𝘤𝘦 𝘰𝘳 𝘧𝘢𝘪𝘭 𝘵𝘰 𝘥𝘦𝘵𝘦𝘤𝘵 𝘴𝘢𝘧𝘦𝘵𝘺-𝘳𝘦𝘭𝘦𝘷𝘢𝘯𝘵 𝘧𝘢𝘶𝘭𝘵𝘴. And yet, most safety programs skip this step entirely. At 2Go Solutions, we’ve supported multiple AI safety programs in building qualified tool chains — with the processes, documentation, and evidence needed to meet tool confidence expectations. This isn’t just theoretical. It’s part of building safe and certifiable AI systems. 👉 As a System developer, or lead, what’s your company's practice regarding tool certification? 👉 Also as a tool developer, does your company think about ISO compliance during the development or only after such requests come from OEM's ✍️ 𝗥𝗲𝗮𝗰𝗵 𝗼𝘂𝘁 𝗶𝗳 𝘆𝗼𝘂 𝗵𝗮𝘃𝗲 𝘁𝗼𝗼𝗹-𝗿𝗲𝗹𝗮𝘁𝗲𝗱 𝗰𝗼𝗺𝗽𝗹𝗶𝗮𝗻𝗰𝗲 𝗮𝗻𝗱 𝗰𝗲𝗿𝘁𝗶𝗳𝗶𝗰𝗮𝘁𝗶𝗼𝗻 𝗻𝗲𝗲𝗱𝘀 — 𝗼𝗿 𝘄𝗮𝗻𝘁 𝘁𝗼 𝗮𝘂𝗱𝗶𝘁 𝘆𝗼𝘂𝗿 𝗰𝘂𝗿𝗿𝗲𝗻𝘁 𝗔𝗜 𝘁𝗼𝗼𝗹𝗰𝗵𝗮𝗶𝗻 𝗳𝗼𝗿 𝗜𝗦𝗢 𝟮𝟲𝟮𝟲𝟮 𝗿𝗲𝗮𝗱𝗶𝗻𝗲𝘀𝘀. #ISO26262 #FunctionalSafety #AISafety #ToolCertification #AutomotiveAI #2GoSolutions #MLSafety

82

28 Comments

The best place in any office is its library. 📚 Books always remind me to keep growing and evolving. Recently, I’ve been diving into Harvard Business Review articles — they’re a goldmine of deep insights into technology and market trends, from the automotive sector to warehouse innovations. It’s fascinating to see how BMW’s China market is advancing with AI-driven strategies, and how Amazon leverages thousands of LLMs to power its catalog systems and continuously refine user interfaces. Learning never stops — and the library is where innovation begins. #reader #leader

5

Enterprise integration is going through a meaningful shift. For many years, organizations connected partners and internal systems through batch-based integrations and point to point data exchange. While this model served well, today’s businesses increasingly require real-time visibility, faster decisions, and more responsive supply chains. Having spent close to two decades working across EDI, B2B integrations, and large partner ecosystems, I recently completed the Apache Kafka Fundamentals Accreditation from Confluent to further explore how modern event-driven architectures are enabling this transition. What is becoming clear across many enterprises is the emergence of a new integration pattern — where systems like SAP/ERP, WMS, and TMS platforms are connected through event streams rather than only traditional point-to-point integrations. For example: An order created in SAP/ERP can instantly trigger downstream events that update warehouse operations (WMS), transportation planning (TMS), analytics platforms and customer visibility systems — enabling organizations to move from delayed updates to real-time operational awareness. For leaders responsible for integration and platform strategy, the opportunity is to build on existing B2B and API foundations while enabling event-driven enterprise architectures that support modern digital operations. Looking forward to continuing conversations around how organizations are modernizing integration platforms to support real-time, connected business ecosystems. #EnterpriseArchitecture #DigitalTransformation #EnterpriseIntegration #SupplyChain #Leadership #Technology

3

Auto OEMs are incurring significant costs by repeatedly rebuilding the same data pipelines across various teams, including Engineering, Connected Vehicle, After-sales, Warranty, and Manufacturing. This leads to: - Data silos - Duplicated ETL processes - Increased cloud expenses A mindset shift is essential: treat data like a product, not a project. By establishing a Data Exchange Platform with the following features, costs can be significantly reduced: - Standard schemas - Common ingestion - Shared storage - Role-based access - Consumers subscribing to data instead of rebuilding pipelines This approach can lead to: - 40–60% reduction in pipeline development costs - 30–35% savings on cloud expenses - 70% faster onboarding of new signals and sources Additionally, AI enhances this process by: - Auto-mapping datasets - Auto-generating transformation code - Deduplicating pipelines - Predicting data quality issues The future for carmakers lies in a unified platform with multiple producers and consumers, eliminating duplication. Reach out to our team Tata Elxsi for solutions to build your platform for tomorrow. #AutomotiveIndustry #ConnectedVehicles #DataPlatforms #DigitalTransformation #AI

17

2 Comments

As we all know, Salesforce is bringing SFMC Next, where Salesforce Data Cloud becomes the primary data source for activation. However, Data Cloud does not provide the same level of flexibility that traditional SFMC Engagement offers—especially around heavy SQL processing and Automation Studio–based fulfilment logic. Has anyone started migrating from Marketing Cloud Engagement to SFMC Next? If so, how are you handling complex SQL-driven fulfilment logic and automations that were previously built in SFMC but are not natively supported in Data Cloud? Interested in hearing your experiences. #Salesforce #SFMC #SFMCNext #SalesforceDataCloud #MarketingCloud #MarTech #EnterpriseArchitecture #SolutionArchitecture #CRM

22

Turning Vibe Coding Into a Live App: Token Length Checker Plus (Gemini) 🚀 Just shipped — a Gemini-powered web app that helps in: ✂️ Summarize long text 🧩 Simplify complex paragraphs 🔎 Extract Keywords & Hashtags 🧮 Estimate ≤ token length (great for LLM prompts/posts) 📝 Use an interactive quiz card to self-check understanding GitHub Repo: https://lnkd.in/gduSMZ7K 💡 Key learnings Spinning up an Ubuntu EC2 instance and running a Flask app on port 5000 Setting up a clean virtual environment and installing dependencies Configuring security groups so the app is reachable during testing Using system packages effectively (Python, pip, basic Linux tooling) 🌟 How it can be useful Students & teachers: Turn dense notes into quick summaries; quiz yourself to retain more Creators & marketers: Pull keywords/hashtags instantly and make content easier to read Engineers & PMs: Trim specs into concise briefs; keep drafts under token limits Anyone writing online: Get fast, clear versions before sharing/posting 🎯 Motivation & use I wanted a simple, everyday writing helper that’s fast, clean, and modular: Card-based UI with an interactive quiz, keeping raw JSON out of the way Add-on style architecture so features evolve without cluttering the core 🙏 Special thanks to Rohan Shravan for giving me the push to actually post on LinkedIn  #Flask #Python #Gemini #LLM #AWS #EC2 #LearningInPublic P.S : 🚧 Work in progress: polishing the UI and adding small UX tweaks next. ✨ Iterating fast—this is v1; more refinements coming soon. 🛠️ Still evolving: open to feedback before I lock v2.

29