Big Idea

The argument in one line.

Building a productive AI operating system requires starting with a detailed PRD that maps your data architecture, projects, and skills before implementation, not after.

Who This Is For

Read if. Skip if.

READ IF YOU ARE…

A solopreneur or small founder who uses Claude regularly and wants to systematize repetitive work like daily digests, research, and project tracking.
Someone building a personal knowledge management system who's comfortable with prompt engineering and wants copy-paste templates they can adapt immediately.
A creator or operator who already uses AI tools but feels disorganized and wants a specific folder/dashboard architecture to scale their workflows.

SKIP IF…

You're not currently a Claude user or have no immediate use case for autonomous AI agents building software overnight.
You're looking for a no-code, plug-and-play solution — this requires writing and iterating on custom prompts and system instructions.

TL;DR

The full version, fast.

A personal AI operating system works only when discipline comes before code: every project begins with a product requirement document that defines the problem, scope, constraints, and success criteria, and nothing gets built until you sign off. The architecture is a data-lake pattern � one local folder ingests fresh data from calendar, email, chat, news, and investments, while dashboards, a daily digest, and task-specific skills sit on top and transform that data into views and actions. An autonomous overnight builder reads pending PRDs on a schedule, ships them while you sleep, and logs results. The lesson: enforce PRD-first behavior, force the agent to push back, take notes aggressively, and confirm anything irreversible.

Members feature

Chat with this breakdown.

Modern Creator members can chat with any breakdown — ask for the hook, quote a framework, find the exact transcript moment. Unlocks at T2: refer 3 friends + add your own API key.

Create a free account →

Chapters

Where the time goes.

00:00 – 00:43

01 · Cold open — proof before tutorial

Demo reel of the finished system: morning digest, investments dashboard, mission control showing overnight builds. Promise stated: copy-paste prompts provided, no coding required.

00:43 – 02:49

02 · Step 1 — Operating instructions

Paste a master system prompt into Claude Cowork settings. Four non-standard sections: PRD First Always, Pushback and Clarification, Note Taking, Reversibility.

02:49 – 03:41

03 · Step 2 — Write the initial PRD

Called the most crucial part. Blueprint analogy. Shows the actual mission-control PRD: projects, architecture, hour-by-hour build plan.

03:41 – 06:54

04 · PRD deep-dive — projects + architecture

Projects: investments dashboard, morning brief, custom skills (today/research-ticker/deep-prep), autonomous builder. Three-layer folder system. Data-lake analogy for data flows.

06:54 – 09:19

05 · Sponsor — Hostinger VPS

Claude Cowork dies when laptop sleeps. Hostinger KVM 2 VPS keeps agent running 24/7. Code: tina10.

09:19 – 10:27

06 · Step 3 — Set up folders and project

One-minute live demo: create folder, open as Cowork project, paste instructions, drop in PRD, type start command.

10:27 – 12:47

07 · Step 4 — Pure execution

Hours 1-4 of build. Shows live screenshots: investments dashboard, morning digest output, prep skill output. Notes 95% smooth with a PRD.

12:47 – 14:09

08 · Step 5 — Autonomous overnight builder

Pending/in-progress/done/failed folder workflow. Scheduled task every 30 min scans /pending and builds. Mission control dashboard tracks all builds.

14:09 – 15:23

09 · Wrap-up — memory + quiz + CTA

Memory bloats over time. Teases part three on complex memory systems. On-screen quiz + subscribe CTA.

Atomic Insights

Lines worth screenshotting.

Requiring Claude to write a PRD before building anything prevents the classic 30-minute drift where it builds something entirely different from what you wanted.
A data-lake architecture — pipelines feeding a central folder — is the infrastructure pattern that makes every AI-built dashboard actually useful.
An autonomous overnight builder that processes a pending folder every 30 minutes lets you wake up to finished software instead of a to-do list.
Operating instructions that explicitly tell Claude to push back and flag trade-offs produce better output than instructions that just ask it to comply.
The local AI agent problem is that the moment your laptop sleeps, hours of agent progress vanish with it.
Asking Claude to document and take notes aggressively as part of its core personality — not as an afterthought — is the only way memory stays useful over time.
Claude Cowork memory starts degrading over time, so a formal memory management system is not optional for power users.
A three-layer data architecture separates raw data, processed outputs, and project builds — without that separation, everything becomes a mess quickly.
Forcing explicit sign-off before any build starts is cheaper than discovering an hour later that Claude misunderstood the scope.
Building dashboards first is the fastest way to make an AI system feel real — visibility into what's happening beats blindly trusting logs.
A morning digest that pulls calendar, email, Slack, and investments into one view is buildable by a non-developer in a single five-hour session.
Scheduling tasks every 30 minutes to check a pending folder gives you asynchronous, fire-and-forget delegation to an AI agent.
PRDs aren't bureaucracy — they're the blueprint that determines whether what Claude builds is what you actually wanted.
Paying for a max-tier AI subscription is only worth it if you build systems that run autonomously while you sleep.

Takeaway

The PRD is the product.

Build discipline, not just build speed

Tina's system works not because Claude Cowork is magic, but because she made Claude commit in writing before touching a line of code — and that discipline is portable to any tool.

Add a PRD-First rule to any agent system prompt: before building anything, write a PRD with problem, success criteria, scope, and build plan, then ask for sign-off.
The pending/in-progress/done folder pattern is simpler than any UI — adopt it for JoeFlow Sessions batch execution instead of building a live trigger board.
The four operating-instruction sections (PRD, Pushback, Notes, Reversibility) are a template droppable into any project CLAUDE.md today.
Proof-first video structure — show finished system in first 40 seconds — is what the JoeFlow sales page and demo reels are missing.
The data-lake framing makes AI architectures explainable to non-technical audiences: steal it for MCN+ onboarding copy.

Glossary

Terms worth knowing.

Claude Cowork: A local AI agent from Anthropic that runs on your computer, organizes work into projects and folders, and can execute multi-step tasks like building dashboards, running skills, and autonomously generating software.
PRD: A product requirements document that spells out the problem, success criteria, scope, constraints, and build plan for a project before any code is written, so the builder knows exactly what to ship.
Skill: A reusable set of instructions that tells an AI agent how to run a specific recurring task, such as fetching a morning brief or researching a stock, triggered by a slash command.
Meta prompt: A prompt whose job is to generate another prompt, typically by interviewing the user and producing a customized instruction set or document tailored to their context.
Data lake: A central store that pulls in raw data from many sources (calendars, emails, finances, news) so other tools can query, transform, and present it without each app fetching its own copy.
Daily digest: An automated morning briefing that compiles calendar events, emails, chats, investments, and news into a single summary delivered each day.
Mission control dashboard: A single-pane interface that surfaces the status of ongoing builds, scheduled jobs, and project health so the operator can monitor an autonomous system at a glance.
Autonomous builder: An agent loop that watches a folder for new project briefs, picks them up on a schedule, builds the software unattended, and moves each job through pending, in-progress, done, and failed states.
Connectors: Integrations that let an AI agent read from and write to external services such as Gmail, Google Calendar, Slack, or brokerage accounts.
Plug-ins: Optional add-on modules that extend an AI agent with extra capabilities, commands, or workflows beyond what ships in the base install.
Memory system: A structured way for an AI agent to record notes, decisions, and context across sessions so it can stay coherent on long-running projects instead of forgetting prior work.
Scaffolding: The initial folder structure, config files, and placeholder documents that have to exist before an agent can start filling in real work on a project.
Scheduled task: A job set to run automatically at a fixed interval, used here to have the agent check a folder every thirty minutes for new briefs to build.
VPS: A virtual private server, a rented always-on cloud machine that runs your software around the clock so an agent's work continues even when your laptop is closed.
Anthropic API key: A private credential that authenticates programmatic access to Anthropic's Claude models and is required to run Claude-powered agents outside the consumer chat app.
Claude Max subscription: Anthropic's highest-tier consumer plan, which raises usage limits enough to support heavy power-user workflows like running multiple agents and long autonomous builds.
OpenCode: An open-source local coding agent, mentioned as a peer to Claude Cowork that shares the same problem of losing progress when the host machine sleeps.
Reversibility: A working principle that an agent should pause and confirm before any action that is hard or impossible to undo, such as deleting files or sending messages.

Resources Mentioned

Things they pointed at.

00:43toolClaude Cowork ↗

04:06channelClaude Cowork fundamentals video (Tina Huang channel)

08:14productHostinger KVM 2 VPS ↗

13:36toolOpenClaw

Quotables

Lines you could clip.

02:44

“Interrupting with a question is always cheaper than silently destroying something.”

Self-contained principle, zero setup needed→ TikTok hook or IG reel cold open↗ Tweet quote

02:57

“You can technically YOLO a blueprint that's not very good and then your building is also gonna fall down.”

Funny analogy that lands the PRD argument in under 10 seconds→ TikTok hook↗ Tweet quote

05:05

“I woke up to it done. Very delightful.”

Tight, visual payoff — the whole system in five words→ IG reel cold open↗ Tweet quote

12:39

“I pay for a Claude Max subscription. I gotta get all the use out of my tokens too.”

Relatable power-user admission with deadpan delivery→ Newsletter pull-quote↗ Tweet quote

The Script

Word for word.

metaphoranalogy

00:00This is my Claude Corex setup. Every morning, I receive a daily digest on my Apple notes with all of my investments, calendar events, emails, and action items for the day. And if I wanna dive deeper into my investments, I can go to my custom investments dashboard.

00:13I also have co work autonomously building me new software and workflows every night. So I can also check out my mission control dashboard and see what projects it's built for me. For example, last night, it built for me this resource hub.

00:23So in this video, I'm gonna show you how to set up a system like this, but custom for you. I will also be giving you my co work prompts along the way, so you can just copy paste and steal them directly. You're welcome.

00:33And, no, you don't even need to know how to code. Anyways, without further ado, let's go. A portion of this video is sponsored by Hostinger.

00:43Step number one is configuring your settings. Of course, if you haven't already, please do download Claude CoWork. And after that, you'll be greeted with the Claude hub.

00:51Now you wanna go to Claude settings and then click co work and paste in your operating instructions.

00:59I'm gonna put a prompt on screen and also in description that you can paste into the chatbot tab of Claude Co work in order to get this meta prompt. However, I do wanna draw your attention to some parts of this instruction because it's a little bit nonstandard, but trust me, you would get so much more out of Cowork with this.

01:13Okay. So this stuff is pretty standard. Blah blah blah.

01:15You're Claude working as Tina's primary AI hub in Cowork. Then there are four sections. So building things, pushback and clarification, note taking, and reversibility.

01:25Since I use CoWork to build a lot of stuff, we're gonna be using CoWork to build a lot of stuff. So the most important thing here is PRD first always. I'm basically asking CoWork that before it builds anything, it has to write up a PRD, a product requirement documents that list out in detail what it is that it's going to be building.

01:40Like this, for example, it includes articulating what the problem is, what's the success criteria, what is the scope of what we're building, constraints, build plans, etcetera. This is really important to guarantee that cohort is actually building what you want it to be building. So you won't be surprised, like, I don't know, like, thirty minutes, an hour later, it builds something, and it's, like, not even what you want.

01:57So, yes, I will always ask you to have a PRD, also to ask me any open questions to make sure that coworker understands what it's building, and finally, to ask for explicit sign off before building. I also have a section on pushback and clarification because I want Claude to push back on me, question what it is that I'm doing, and to disagree when the plan seems off strategy, technically wrong, or inconsistent with prior decisions.

02:18Also to flag trade offs that I may not have considered because I am not infallible. Then there is note taking. I tell co work that it's to take notes aggressively.

02:26Memory management is really, really important when it comes to co work, and I'll be talking more about it later in the video. But I do make it as a part of its personality from the very beginning to always document everything. Then finally, there's reversibility.

02:35This, I'm just asking you to always confirm before doing anything that is hard to reverse. If in doubt, stop and ask. Interrupting with a question is always cheaper than silently destroying something.

02:45Great. Now that is complete. Let's move on to step two, writing the initial PRD.

02:53Okay. Awaken if you are falling asleep because this is the most crucial part of this entire build.

03:00Okay? Focus. The analogy I'm gonna make here is kinda like coming up with a blueprint of a building that you're gonna build.

03:06Because you can technically, like, YOLO, make a blueprint that's, like, not very good, and then you're probably your building is also gonna be not very good and, like, fall down or something. Right? Very bad.

03:14So this PRD is kinda like the blueprint of how you're gonna set up your core system and all the projects that you're gonna be building now and henceforth. I hope I used that word right. Henceforth.

03:23Uh, from this time on or from that time on. Yes. And henceforth, what this PRD looks like.

03:29Okay. So mission control build PRD. It contains everything that you need in order to set up your initial core workspace and initial projects.

03:36As you can see, it is very long and very detailed. So luckily for you, I'm going to link a prompt below that you can copy paste into your Claw Chat.

03:46It will ask you some questions and then generate your version of this PRD custom to you. Man, I am so jealous of you that I'm giving you this. Okay?

03:54Here, I do wanna go over some crucial parts of this PRD as well because it's gonna be important for the build later. So the first step is to figure out what projects you want core to be building. In my case, to start off with, I wanted to have a dashboard, specifically an investment dashboard where I can track my investments, update them, and to be able to get news and custom advice based upon my investing philosophy.

04:13I also wanted to build a daily digest, a morning brief that will give me relevant updates across my Google Calendar, my emails, my Slack chats, and relevant news. I also want to have some custom skills, which are set instructions that tell Claude how to run a specific task. By the way, this is the first time you heard about core skills.

04:30I highly recommend that you check out this video over here, which is, like, the fundamentals, the prequel to this video where I introduce you to core, explain how it works, and all its functionalities, including things like skills in a lot of detail. So I'm not gonna reexplain all of this in this video. If you are a long term viewer of this channel, you will know that I always insist that you learn the fundamentals of things because it's gonna make you so much better at using tools and building things.

04:51Anyways, done with that. Check out that video if you haven't already. And going back to skills I wanna start building, specifically, I wanted to start off by building three different skills.

04:59The today's skill where I'm able to pull up the morning brief again and update it. The research skill that will deep dive into a specific stock I'm interested in and deep prep skill for the meetings that I have. Because oftentimes, I don't really know the contacts for the meetings I'm going into.

05:13A lot of people have access to my calendar, so they just drop stuff into my calendars, and sometimes I don't even know who I'm gonna be meeting with. So it helps a lot. And finally, the autonomous builder.

05:21This allows me to drop a brief, a PRD of something that I want to build. And for Cloud to autonomously build it usually through the night so I can wake up to a finished work project and be very delighted. Like this resource hub that I woke up to today.

05:33Very delightful. These are just projects that I want to start up building. Of course, I've built a lot more things as well.

05:38And for you, it's probably gonna be different. So if you go through the prompt to generate your own PRD, it will help you to figure out the starter projects that are best for you. Okay.

05:45So after you figure out what kind of projects that you want, another section that is very important is the architecture overview. You see, CoWork is a local AI agent, which means that it has to operate on your local computer, your local machine. And in my case, it operates under a folder called a CoWork that has a bunch of things in it.

06:00There are actually a lot of different ways that you can set up your folder system. Not gonna go into more too much more detail here Check out the fundamentals video. But just know that in my case, I have a three layer system, a corresponding memory architecture, and the PRD actually maps out what the folder structure should look like here.

06:16This is a little bit abstract, so don't worry so much about it. We'll be building this out later in the video, just wanted to call your attention to this part of the PRD. Final part I wanna draw your attention to in the PRD is the deep build plan.

06:27So I told a coworker that I had five hours, um, to build out all these things in the prompt. Of course, when you go through the prompt that I gave you, you can adjust the hours to however many hours that you want. But, yes, for mine, I had five hours, and it laid out a plan hour by hour for me.

06:41The way it works is that hour one is setting up the data layer foundation. You can think about this like if you wanted to build a lake, a man made lake. You have a big hole, but you gotta fill it with water.

06:52So you need to set up different pipelines for water to trickle into the lake and for it to be constantly moving. Now in this analogy, the water here is data.

07:02You need to have different pipelines in which you're getting fresh data that's gonna be going into your lake and moving along. So the folder structure and different tasks, which we'll talk about later, is how you're gonna set up this lake. Then in hours two, three, and four, you're building things on top of this data lake.

07:18You could be building, like, a dam, a water wheel. What other things do we build on lakes? A walkway, a fishing section.

07:25Yes. Those will be, like, your projects, like the investment mission control, the morning brief, and your autonomous builder. Because all of these things that you're building is basically just taking the data and then transforming them, doing stuff to it, we're presenting it in a different way.

07:38Like, any type of dashboard, like investment dashboard is taking your investment and financial data and then portraying it. Right? A morning brief is taking data from, like, your your calendars, your chats, emails, investments, and compiling them together and presenting it to you.

07:50A general builder is transforming a brief, a PRD into actual software, etcetera, etcetera. I hope that makes sense. If you think about your projects and your workflows in this way, I think it becomes a lot more clear about how it is that you should be building them out.

08:02Great. And let's actually get things started by moving on to step three, setting up folders and project. The new launch of Cloudco remote is pretty amazing, but the second your laptop sleeps, your agent dies with it.

08:13The hours and hours of progress on a refraction or migration are just gone. By the way, this is the same problem with all local AI agents like OpenClaw, Hermes, and a co work. Luckily, Hostinger solves this.

08:24Your agent runs on your own private VPS twenty four seven. Your API keys, your code credentials, all on your own infra, never on a third party platform. And because it never sleeps, I can kick off a task on a phone, close the app, go to bed, and wake up to it done.

08:37You can head on over to hostinger.com/tina10 and grab the KVM two plan. It's built for exactly this kind of persistent heavy workload.

08:45I went with a twenty four month plan for the savings, and my code, Tina Huang, gets you an extra 10% off. The setup is genuinely fast. The Cloud Code template is preselected.

08:54You click deploy, authenticate with your Anthropic API key, and voila, you are live. I have this running in minutes. So if you want your own always on Cloud Code agent with zero setup friction and data privacy, use my link host thinker.com/tina10 and apply code Tina Huang for an exclusive discount on the Cloud Code VPS plan.

09:12Link is in the description. Thank you so much Hostinger for sponsoring this portion of the video. Now back to the video.

09:19Alright. So I have a good news and bad news. Which do you want to hear first?

09:23What's that? Bad news first? Okay.

09:25So, basically, before you give this entire PRD to COWERK and just tell it to do it, you do need to set up the initial folder and scaffolding. The good news is that this literally takes one minute. So let me show you.

09:36Okay? So you go to your finder and you create a folder. Already have one called co work, so we'll just call this one mission control.

09:42Then you go to co work, click projects, new project, use an existing folder, and then find your folder, and click open. Then in the instructions, I just put some additional detail referencing the prdmissioncontrol.md that I wanted to follow.

09:57I'll put the instructions in the description as well and click create. Then I go back to my mission control folder, click into it, and drop in the PRD mission control dot m d file. Then you just type, please start building with the mission control PRD.

10:13And that's it. Great. You've set things up now.

10:15You can see that it tells you what the next steps are, like going to personal plug ins, installing plug ins, connectors, and skills. Again, if you don't know what these are, please check out my core fundamentals video.

10:25And, yeah, you basically just follow along.

10:30Now we're on to step four, pure execution. So you basically just follow co work and go through the detailed PRD plan, starting off with the folder structure, making sure to do slash start from the productivity plug in to kick off the memory system, finishing up hour one, building your lake, making sure all the water, the data is flowing in.

10:48Then in hour two, you're building the investment mission control dashboard scoped out here in the PRD, and it looks like this. There are, of course, a lot of other dashboards that you can build too. Like, here's my mission control dashboard where I can look at different builds and schedules, things like that.

11:02Dashboard is one of my favorite things to build whenever I do a project because it allows me to visualize things and monitor what's going on. I'm also gonna put on screen some other examples of dashboards that I have built and things that you could consider building too. So I do wanna make a caveat here.

11:14It's not gonna be, like, completely smooth sailing. Having a PRD and following through it is gonna get you, like, 95% of the way, but you do need to be doing things like giving permissions and potentially making some tweaks and additions to the PRD as you're building along as well. Honestly, it's pretty smooth.

11:27Then hour three is going to be the morning breathe and the skill suite. Here's the scope of it in the PRD. This is what it looks like.

11:35It gives me an overview of my portfolio, what happened during that, tells me about my day, my inbox, my calendar, things I need to follow-up on, and tells me what I'm gonna be doing for the day. This daily digest is custom to me, so I'm gonna put on screen now some other things that you can put onto your daily digest if you wish.

11:51The today skill is able to fetch the on demand version of the morning brief if I want to look at it or update it. The research ticker skill that allows me to do slash research whatever stock it is I want to deep dive into and then gives me information about it, which I can also see updated on my investments dashboard. And the prep skill where I can input any meeting like next week's team meeting, and it would tell me all the context about the meeting and the people in the meeting.

12:15There are other skills I have built and things that I plan to build as well, so I'm gonna put all of those on screen as inspo. Amazing. So by hour three, I had built out all of these.

12:23And the final thing is the general builder. So I consider this to be, like, an advanced bonus section, so please feel free to skip it. But I included even my initial build because I just really like autonomously building things.

12:36I pay for a Cloudmax subscriptions. I gotta, like, get all the use out of my tokens too. You know?

12:40That's how I justify it. So that will be step five, the autonomous build.

12:47So the way that you build this is actually still the same. Here's the scope of it, including the mission control PRD. You just follow through with instructions, and it should build up for you.

12:54But I do actually wanna demo this for you because it's a little bit more complex than the other things. What this is is that I will have certain task workflows and softwares that I want to build across my company, my personal life, my investments, etcetera. I asked Kory to give some suggestions of different things that we can be building in order to improve this system and the business and my life.

13:12It will draft the PRDs. I will go through an approval with it, make sure these are the things that I want to build, and then it will put it into the pending folder. Then I have a scheduled task where every thirty minutes, it will check the pending folder, and then it will pick it up.

13:25And if it sees that there's a PRD, then it would go and build it. By the way, not to be super complicated here, I actually do also give some tasks, some PRDs to Cloud Code, uh, depending on how advanced the project is, but I'm not gonna go into more detail about that. Anyways, while it's building it, it would put it in the in progress folder.

13:42And then when it's done, it would put it into the done folder. Of course, if it failed, it'll put into the failed folder, um, and it also keeps logs on what it is that it's built. I do also have this mission control dashboard where I can see the builds more clearly.

13:55Just made a dashboard for us so I can more clearly visualize things. And, yeah, I've built a bunch of different things in this way. I will put on screen some of the things that I've built, some of the things that I want to build as well, um, and other suggestions that you can take as inspo if you want to implement this more advanced autonomous builder section.

14:10Really cool. Right? I love it.

14:12I really love waking up to something new that it builds. Sometimes I use it, sometimes I don't, but a lot of it I actually have used and built on top of them as well. So, yes, let me know in the comments if you're gonna implement this section.

14:22Kinda curious to see if other people are into this too. That was hour four, and hour five was just to polish things up, like having notification set up and the end to end testing. So not gonna show you that part pretty mundane.

14:32But, yes, it was actually pretty accurate. It really did take me around five hours to get this done. Of course, afterwards, I built a lot more things on top of it, like, for tracking my health, a lot skills related to planning my day, daily plans, weekly plans, and improvements with memory as well because I just build a lot of stuff.

14:47I am definitely a co work power user. So as time goes on, co work does start forgetting things. It's not as bad as OpenClaw if you've used OpenClaw before, but your memory system does start bloating.

14:56So I've implemented a more complex memory system as well, which I'm not gonna go into too much more detail about this now because it is a little bit more technical, but maybe I can do a part three on Coreg. Let me know in the comments if you would like me to do that. Your wish is my command, mostly, sometimes.

15:12I'm gonna sneak in a little quiz now, which I'm gonna put on screen. Please do answer these questions to help you retain all the information that we cover today because we covered a lot today. And I will see you guys in the next video.

15:21We're livestream.

The Hook

The bait, then the rug-pull.

Before the tutorial starts, Tina shows you the finished product: a morning digest that writes itself, an investments dashboard that updates live, and a mission control screen showing what Claude built overnight. Only once you have seen it working does she explain how any of it was made.

Frameworks

Named ideas worth stealing.

01:32concept

PRD-First Always

Write PRD with problem, success criteria, scope, constraints, build plan
Ask open questions before starting
Require explicit sign-off before building

Claude must write a Product Requirements Document before touching any code. Eliminates the hour-wasted-building-the-wrong-thing problem.

Steal forAny JoeFlow feature build, Mod Boss AI, or Sessions batch task

00:56list

Four Operating Instruction Sections

PRD First — document before building
Pushback — disagree when off-strategy or technically wrong
Note Taking — aggressive documentation as personality trait
Reversibility — confirm before hard-to-reverse actions

System prompt architecture that turns Claude from a yes-machine into an accountable builder.

Steal forJoeFlow Chef context / system prompt for Sessions panel

06:42model

Data Lake Analogy

Hour 1 = dig the lake and set up data pipelines (calendar, email, Slack, news). Hours 2-4 = build on top (dashboards, digests, builders). Every project is just transforming and presenting data.

Steal forExplaining the MCN data architecture to non-technical audiences

12:47model