4 Exercise: Project Progress - Video Tutorials & Practice Problems
Video duration:
5m
Play a video:
Video transcript
<v ->All right, time to get back to your work.</v> What are you gonna do now to make your course project better? So I've got a few things lined up for you. Let's walk through them. So the first thing is if your project needs a data scientist, if you're working maybe on one of those specialized approaches where you're creating your own model, I want you to think about what you would do if they assigned a data scientist to your project tomorrow. What job would you give that person? So for your first initiative, what features do you think might matter? So using your domain expertise, what are the features that you think, what are the characteristics you think, what's the data you think might really help solve the problem? What are you going to want that data scientist to analyze? So he or she is gonna come to the project with all sorts of understanding of AI techniques and data science, but they might not come with a real understanding of your problem and they probably won't understand all the data that might be available. You might know more of that than they do. So the question I have for you is for your project, if you need a data scientist on your project because you are creating some kind of specialized model, what job would you give them? What are the features that you think you should suggest to them for them to start to investigate? Next thing is how would you persuade someone who really isn't ready to take the plunge with Agile? They like Waterfall, they're used to it, they've been doing it since the year three. What are you going to do to help that person? How are you going to overcome their objections? What objections do you think they'll have? What do you think that they'll be worried about? And what would you say to persuade them? How would you help them to overcome their fear? How would you help them overcome their resistance to change? How would you persuade them to give Agile a try? Now, the reason to do that is because Agile is really going to be part of the secret of delivering value from AI. And so examine your current development process within the organization that you've chosen. So if you're already using Agile, then maybe you're fine, although there could still be some issues where they're not using Agile, maybe the best way they could, but especially if you're using Waterfall or other types of really structured processes, try and identify what kind of issues you think will come up as you try to implement your Agile project that this current process really isn't designed for because you might need to get some kind of dispensation from the normal way that you do projects in order to make your project work. So why do you think Agile will help? How are you going to persuade people? What kind of obstacles are in the way of you just taking an Agile approach for your project? What kind of things are going on in the organization that might make that hard? Is there some type of approach that you can take that will overcome that? I mean, a lot of times, especially big companies, have almost a corporate immune system where they resist all sorts of big changes. And so how can you overcome the corporate immune system or avoid it so that you're able to use an Agile process for your project even if everybody else is using Waterfall or other structured techniques for theirs? And how will your first initiative work? So you don't have to draw out some kind of diagram like this, but can you think about whether you've got a UI that you have to make some strategic decisions on? Is there kind of a two-phased process where first you're training your model and then you're using your model in production to make predictions? So how does that work? Where would you get the data from? How would you train that model, and how would you know when it's good enough to put into production? Now we're gonna go deeper on all these kinds of questions, but if you just could kind of sketch out how that might work, that's going to really help you to go deeper as we do these kinds of things in later lessons. And so let's summarize all the things that you're gonna do right now. First, lay out how you would work with a data specialist and suggest some features that you think might be important for your first AI initiative. And I also want you to think about the current process you're using in the organization. How would you brainstorm ways of getting an Agile skeptic to be willing to go along with this new kind of process and what are some ways that Agile will really help if your organization hasn't done it before? And try and speculate on how you could implement your first initiative. I know you're not gonna have all the answers yet. That's okay, and in fact, you may not have the skills to ever have all the answers. That's what you're gonna work with the rest of the team to do. But if you could at least draw out on a back of a napkin what you think the system might do, even as just a starting point for discussion, that's gonna really help you communicate with the other people on the team that you'll need help with.