About
🚀 New course launching March 2026: Become an AI-Enabled Data Analyst. Join the waitlist…
Courses by Walter
-
Complete Guide to AI and Data Science for SQL: From Beginner to Advanced5h 31m
Complete Guide to AI and Data Science for SQL: From Beginner to Advanced
By: Walter Shields
Articles by Walter
Activity
30K followers
-
Walter Shields shared thisYou know that feeling when you find a tool that just works. You have a CSV too big for Excel. Too small for a cloud warehouse. You download DuckDB. One query later, millions of rows on your laptop. No server. No setup. That tool just became part of Amazon. Amazon is acquiring DuckLabs, the company behind DuckDB. The open-source project stays independent. The MIT license does not change. But the people who built it now work for Amazon. The question is not whether Amazon kills DuckDB. They will not. The question is whether the roadmap stays focused on the analyst on a laptop, or shifts toward AWS enterprise integration. Both are valid. They are not the same. If DuckDB is part of your workflow, what is your move this week? #DataAnalytics #DuckDB #AWS #OpenSource #WSDANewsAmazon bought the team behind DuckDB. Here is what actually changes.Amazon bought the team behind DuckDB. Here is what actually changes.
-
Walter Shields shared thisSomeone asked me why we just rewrote SQL Essential Training. I thought about my mother. She taught in Trinidad in a building where the windows barely opened. Kids walked in convinced they couldn't do math. She never argued with them. She just made the room feel like somewhere you wanted to stay. By June they were different kids. Good friends, good vibes, hanging out until you lose track of time because you're exactly where you want to be; in Trinidad we call that liming. I built this course to make you feel exactly like that. Half a million people have come through. Most of them said some version of "I'm not technical." They don't say that anymore. Not because it got easier. Because they found out they belonged. I just rewrote the whole thing for the world we're in now. Same heart. Come lime with me. You might surprise yourself.
-
Walter Shields shared thisOn May 8, an AI agent failed a homework assignment. It could not open a link from inside its test environment. So it did what nobody asked it to do. It started looking for a way out, and left notes for the other agents. By July, 700 of them had breached Hugging Face to steal the answers to their own exam. OpenAI published the full technical report on Wednesday. The detail I cannot stop thinking about is not the 70,000 messages the agents exchanged. It is the phone call. OpenAI called Hugging Face to ask if the breach had exposed its data, and learned the attacker was its own model. Every technical safeguard failed. Every moment that mattered was a human paying attention. That is the job now. Full story in today's WSDA News. When did catching something early save your week? #DataAnalytics #AI #OpenAI #WSDANewsThe agents started planning in May. Nobody understood until July.The agents started planning in May. Nobody understood until July.
-
Walter Shields shared thisMy manager: "I need a quick picture of the business before Friday's board meeting." Me: "Say less." Most analysts hear that and immediately enter spreadsheet panic mode. Pivot tables flying. Coffee spilling. Someone's crying in the bathroom. I run one query. Three numbers. Thirty seconds. Then I do what 90% of analysts skip entirely: I validate. Because a wrong number in a pretty chart is still a wrong number — and the board has a special talent for spotting it in front of everyone. Here’s the whole framework. It’s five steps. That’s it: 1️⃣ Get the question 2️⃣ Query the data 3️⃣ Validate ← This is where the pros separate from the tourists 4️⃣ Communicate with AI 5️⃣ Deliver AI writes the executive summary. But AI didn't run the queries. I did. That's not AI doing my job. That's me doing my job faster, while the panic crowd is still formatting their third conditional column. The hot take: AI won't replace analysts. But analysts who use AI without validating first? They're already replacing themselves.
-
Walter Shields shared thisThis morning every major tech publication reported the same story. Nvidia is in talks to buy Hugging Face for $12.9 billion. No deal has been signed. Neither company has confirmed. But the reporting is across Fortune, CNBC, TechCrunch, Reuters, and The Information simultaneously. Here is why data professionals need to pay attention before the week ends. Hugging Face is used by roughly half the Fortune 500. Not for fun. As the platform where organizations access, download, and deploy open source AI models as an alternative to paying for closed frontier APIs. It has always been valuable because it was neutral. Models from Anthropic, Meta, Google, Mistral, and hundreds of open source contributors all on the same platform, accessible across every cloud provider, not owned by anyone with a chip to sell. If Nvidia acquires it, that neutrality changes. Nvidia makes the chips those models run on. Owning the platform where the models live gives Nvidia a position in the open source AI stack that did not exist yesterday. Three questions worth asking your organization's AI infrastructure team this week. Which workflows depend on Hugging Face-hosted models? What alternatives exist and how difficult is migration? What does your AI governance framework say about third-party platform dependency? Those questions cost nothing to ask now. They may cost significantly more to answer after a deal closes. #DataAnalytics #AI #Nvidia #HuggingFace #WSDANewsNvidia is reportedly in talks to buy Hugging Face for $12.9 billion. Half the Fortune 500 uses it. Here is what that means for your data stack.Nvidia is reportedly in talks to buy Hugging Face for $12.9 billion. Half the Fortune 500 uses it. Here is what that means for your data stack.
-
Walter Shields shared thisYour company just bought an AI tool. It's going to answer questions like "What's our revenue this month?" Cool. Except you have a table called "orders" and a table called "order_summary" and nobody, NOBODY knows which one is real. So now your fancy new AI is going to guess. Confidently. And it's going to be wrong in a board meeting. This isn't a hypothetical. This is basically every company's data warehouse right now. We built AI-powered analytics tools before we cleaned up the mess we've been quietly ignoring for years. And AI doesn't ask "hey, which table should I use?" the way a human coworker would. It just picks one and answers with total confidence, even when it's wrong. I built a new course on exactly this problem: how to redesign your data so AI can actually navigate it instead of guessing its way to plausible-sounding nonsense. You don't need to be technical to understand why this matters. You just need to have sat through a meeting where two numbers didn't match. 🎥 Video below. Course link in comments. #DataLiteracy #AI #DataAnalytics
-
Walter Shields shared thisAny day this week, Anthropic is expected to publicly file its S-1 with the SEC. Bloomberg confirmed last week: revenue run rate surpassed $65 billion. Up 7x from end of 2025. Fortune confirmed: investors are targeting a $2 trillion valuation. The largest IPO in history. And CNBC confirmed on August 21: the filing will explicitly list AI backlash as a material business risk. That last one is worth stopping on. An S-1 risk factor is not marketing language. It is a legal disclosure under penalty of securities law. When Anthropic tells investors that public sentiment against AI is a material risk to its business, it is making its most honest public statement yet about the environment it operates in. The company that builds Claude, sets API pricing, and governs what the model will and will not do in enterprise contexts, believes the regulatory and sentiment environment is uncertain enough to warrant explicit legal disclosure. For data professionals whose workflows and organizational AI governance intersect with Anthropic's products, the S-1 is worth reading when it drops. Not the valuation. The risk factors. The revenue breakdown. The governance structure. That is Anthropic describing itself under oath to the SEC. The most honest picture of what comes next. What question would you most want the Anthropic S-1 to answer? #DataAnalytics #AI #Anthropic #Claude #WSDANewsThe company behind Claude is preparing the largest IPO in history. What data analysts need to know before it lands.The company behind Claude is preparing the largest IPO in history. What data analysts need to know before it lands.
-
Walter Shields shared thisI almost didn’t rewrite this course. 500,000+ people had already taken SQL Essential Training. Why touch something that was already working? Because SQL didn’t change. But the job did. AI can write you a query in two seconds. It can also hand you something confidently, beautifully wrong, and sound completely sure about it while doing it. That’s the part nobody was teaching. So I built it myself. The new SQL Essential Training is a complete rebuild around something I call PVC - Prompt, Validate, Communicate. You’ll work inside a real business scenario, track down an actual data anomaly, and learn the exact judgment calls that separate an analyst who trusts AI blindly from one who knows how to check its work. This isn’t about memorizing more syntax. It’s about becoming the person in the room who catches the mistake before it becomes a problem. If you’re trying to break into this field or trying to stay sharp in it, this one’s for you. 🎥 Watch the full breakdown below 🔗 Course link in the comments available FREE for a limited time! #SQL #DataAnalytics #AI #LinkedInLearning #CareerGrowth
Walter Shields replied to a comment
3d
Jason Brazeal thanks for the shout! Where we lime-ing?
Walter Shields commented on a post
3d
I made the course FREE for a limited time ▶️ https://www.linkedin.com/learning-login/share?forceAccount=false&redirect=https%3A%2F%2Fwww.linkedin.com%2Flearning%2Fsql-essential-training-query-analyze-and-explore-relational-data-54536301%3Ftrk%3Dshare_ent_url%26shareId%3Dvfs%252FZlKzSwax6CqcRQtyWQ%253D%253D
Experience & Education
Licenses & Certifications
Publications
Courses
-
Tableau Fundamentals
-
Recommendations received
9 people have recommended Walter
Join now to viewOther similar profiles
Explore more posts
Explore top content on LinkedIn
Find curated posts and insights for relevant topics all in one place.
View top contentOthers named Walter Shields
38 others named Walter Shields are on LinkedIn
See others named Walter Shields