Any self-taught AI programmers on IH?
I'm currently in the process of building out my MVP in which I would like to incorporate some AI to facilitate the decision-making process when end-users utilize my product. Given that I'm trying to keep my expenses to the minimal, are there any useful resources out there that I can leverage and better understand exactly what part of AI I'm looking to incorporate into my product? AI is such a broad term, so I'm unsure where to even start.
For context, I'm a data engineer with extensive experience in Python.
Trending on Indie Hackers
I took like 3 coding classes in college. They were not helpful past understanding loops, if then etc.
I am mostly self taught. The biggest thing I learned is pick a small project, cut it in half, then just try to build it. Trying to building something real means you learn so much more than just reading tutorials.
Amen
Thanks for this! I think once I figure out if this is a path I want to pursue for my product at any point in the future, then I'll definitely try out a side project to gain familiarity with a particular concept. And I agree - I don't really grasp the idea just from reading tutorials, and I find it so much more useful building watered down versions of the concepts and building up from there.
The first few lessons of https://www.fast.ai/ may give you the vocabulary to get next steps.
Awesome, this is super helpful and will guide me to figure out (1) if I actually want AI/machine learning in my product and (2) if so, which concepts. Thanks for this!
As you said, AI is a big umbrella term covering many different algorithms in different fields. What you need really depends on your use case. If I were you, I would ask myself:
I'm happy to chat with you and help you figure out what you need. Feel free to reach me at [email protected].
I'll definitely reach out to you by end of today - I want to make sure I know what aspects of AI/machine learning to research to make sure my MVP is set up to incorporate any needed concepts in the future. I'm not sure I want any of it to be within scope of the MVP, but I definitely don't want to re-design the back-end just to incorporate some of it in future iterations. Thanks for the insight!
I am a self-taught data scientist working on several ML algorithms, and also prototyping some user-facing products on the side.
the way you mentioned about decision making at the end-user level itself is very broad to understand. It would be helpful if you put some more context about the problem you are solving for the end-users.
Sure. So basically, I want to help end-users determine what they should negotiate for in a counter job offer, which I'm thinking involves (1) a decision tree, (2) chat bot to help provide assistance if they seem to need help, and (3) some back-end algorithms to draft an output for them based on the decisions made. I hope that helps, but please let me know otherwise! I'm trying not to air out the entire product on here :)
It depends on what types of decisions you want to make but perhaps you could first look into decision trees as they are quite easy to understand and work with. Later on you can of course try harder stuff.
That's what I'm thinking for the MVP. I hoping to learn about AI/machine learning so that I can best structure the MVP to incorporate any potential AI/machine learning concepts in future iterations instead of having to completely re-design the product to fit it in, if that makes sense.
I had AI as one of the subjects in my engineering course. It was extremely boring and theoretical. I picked up AI years later through a few projects I worked on in my workplace. I continued studying AI out of personal interest later. It's vast and never-ending but if you know exactly what you want to achieve you can narrow down to a few standard algorithms or a group of algorithms to program your system to mimic intelligent behaviour.
Coming to your problem statement, you seem to want to find patterns in your end-users behaviour to make decisions, right? Define your problem statement clearly and also what you want your system to do. See if Bayesian Decision algorithms work. https://en.wikipedia.org/wiki/Bayesian_decision_theory
Also, take a look at Markow decision process. We used Markow Chains in one of our product to predict user journeys in our application.
The choice of algorithm will depend a lot on your use case, AI is a huge umbrella term.
Hey Caitlin,
All the resources mentioned here are good and python is definitely the easiest way to implement some “AI” into your product. I think what you’re looking for is some simple machine learning (neural networks). It’s pretty simple to understand, especially if you’re a data engineer.
I have a small SaaS project that uses a neural network to assess the security of a website (all built in python). - I’d be happy to show you how I implemented it!
This comment was deleted 4 years ago.
This is amazing - there are now multiple references to fast.ai so I will 100% take a look there and see if this is something that I even want to have incorporated in my product. I'll send you an email shortly as well!
This comment was deleted 4 years ago.