If you're an experienced coder and are curious about taking Udacity's AI nanodegree, this post is for you.
There were five things that helped me make a decision . I wish future me would have sent me this info back when I was trying to decide. I could have saved lots of time and fear of waisting my money.
Well, future me is here now. I wrote this post in hopes it will help your make your decision to go forward with the Udacity AI nanodegree or not.
If you're interested in getting a sneak peak inside the AI nanodegree, checkout the mind-maps I created. Click here to download them for free.
Here are some things you may have done already or planning to do:
- Did all the research you could and read all the reviews you could find.
- Visited Udacity's website like 20 times looking for something more to help you make the decision.
- Tried to piece your self-learning together. You tried free content online from open course-ware, blog posts and white papers.
- Looked at some books recommended by the experts. But after reading the free preview, you felt a little dumber.
- Keep staring at that $$$ price tag. You can't pull the trigger because its so damn much for an online course.
Well... I did all that and finally broke down and pulled out my credit-card and signed up for the Udacity for AI degree.
The following are 5 reasons you may want to do the same.
1) Self learning AI is not working for you
As developers, we are always in learning mode. The first place we usually start if we what to learn a new piece of tech is google. This usually works well and before you know it you're on your way to incorporating what you learned into a project.
With AI , I spent months reading what I could find for free on the web with very slow progress. It seamed that It kept getting deeper and deeper and got overwhelming real fast. I couldn't seem to wrap my brain around the edges of what all the pieces of AI was and how I could use it.
Here is how taking the AI nanodegree helped me:
How the Udacity AI nanodegree is better than self learning:
- Provided a clear definition of "artificial Intelligence" beyond the trivial wikipedia version
- A structured list or map of all the areas of AI and how they relate.
- Created clarity on how to clarify problems to identify how to solve them with our without AI.
- Because I had to drop big $$$, I got a sense of urgency and commitment. This drove me to carve out time in my week and get it done.
- Focused in on the key areas of AI. There is way too much information out there, I got guidance on whats hot and making the biggest impacts in the field of AI.
As trivial as these points seemed, the nanodegree gave me a new perspective. It provided me understanding beyond what I could price together on my own. It also gave me focus and forced me to commit to learning in a finite time frame.
2) You hate reading white papers
Ok hate is a strong word. I guess the better title would be "You dislike academic, long winded and supper detailed papers when all you want is to get stuff done". Ok, this title is way too long but you get the point.
Many of the advancements in AI are very new and academic. There is very little content available out there for mass consumption. Most sources require prior knowledge in the subject. Or require a deep understanding of math and computer since theory.
Udacity's team did an amazing job. They presented the information as bite sized chunks easy enough for anyone at any level.
Here are some ways Udacity structured the information:
- Simple animated illustrations for complex topics. This was one of the best parts of the content.
- Ease in beginners to reading white papers and only require some reading to complete some projects.
- Introductory lessons on math notations and core CS concepts needed to understand AI white papers.
- The instructor team acknowledged that white papers are currently the best source for latest on AI advancements. The cool thing is that material only referenced white papers as additional reading in most lessons.
Bottom line the Udacity Material is way more beginner friendly than anything I have come across for free on the web .
To get a high level overview of all the topics covered in the Ai nanodegree, checkout the mind-maps I created. Click here to download them for free.
3) You keep looking for the AI hello world example
Many developers are top down learners. They start with an example first, usually the "hello world" example, then dive down into more detailed examples. Developer training is usually structured this way.
Academia likes to start from the bottom up. They like to start with the theory and start building up from there. They then get to the larger picture to encompass generality and complexity to match. Only till the very end of the long winded theory do the academic texts refer to simple examples. Usually, these examples are in quiz form and you're left to search for the answers.
The team at Udacity did well at not being too academic and tried to bring practicality into the fold.
How Udacity was not too academic:
- Although the general structure was still academically oriented and bottom up, they broke theory up into small consumable pieces.
- Lessons were supplemented with practical examples to reinforce what was taught.
- Many of the examples provided, were based on real problems being solved in industry.
The AI nanodegree still feels academic in nature but had enough practical examples to keep me interested.
4) You have a hard time justifying the price
The developer community is is very active online and its not hard to find most information for free. Any structured and paid course usually doesn't go beyond a couple hundred dollars.
The Udacity AI nanodegree is a not a small cost. This was one of the harder objections for me to get over.
The team at Udacity trys to add as much value to the program beyond being another online course.
How Udacity was more than an online course:
- The team at Udacity will assign a personal mentor to give you 1 on 1 answers and guidance.
- A career prep supplement with resume writing and networking tips and best practices.
- Very active Slack community covering industry news, events, project discussion and contests.
- Multiple required projects with code reviews and automated tests with scoring rubrics.
- Life time access to all content with future updates.
Even with all these add-ons, the cost felt a little on the higher end of what I'm comfortable with. Add-ons like the career advice is more targeted to junior candidates. On the other hand, the 1 on 1 mentor was very valuable and something you don't get with typical online courses.
5) You don't feel you meet the prerequisites
When I first read the prerequisites, I got a little intimidated. Even with a good background in math and programming, I haven't done any python yet and I haven't done anything in math since college.
Many developers I talked to said they weren't sure they would be able to hack the math part of learning AI . Some of the best coders I know don't have a strong technical college background if any at all. This was a big sticking point for many devs when I talked to them about learning AI.
I spent tons of time before the first term studying math, python and other CS topics. At the end, I found that 90% of what I studied turned out to be irrelevant.
I still ended up doing just-in-time learning for the areas that I missed in my prep work.
The prerequisites your really need before you take the Udacity AI nanodegree:
- Python basics with focus on looping, recursion and data structures
- Using jupyter notebooks with python
- Condition and non-conditional probability basics.
- Some calculus with focus on derivatives if you really want to get into the theory.
- Very basic linear algebra with focus on matrics multiplication.
The area that slowed me down the most was not the math. The Ai nanodegree covered most of the math in refreshers and intros. I did struggle with figuring out efficient way to manipulate data using python and data libraries like numpy.
Don't get intimidated by the math prerequisites. If you're a good coder you shouldn't have too much difficulty getting through it.
I still would highly recommend doing as much pre learning in all the math areas mentioned. If you want to get the most for your $$$, you need to understand the theory and that requires understanding the math.
So was it worth it?
I had a positive experience with the AI nano-degree and would recommend it to any developer willing to drop the $$$. Go for it if you're looking for the fastest way to learn a whole boat load of AI in a relatively short time. Just like any paid education, you get out of it what you put in. So make sure you prep well and slice out enough time in your week to maximize your learning.
Here are the highlights of the Udacity AI nanodegree:
- Started from basic AI algorithms then moved to very advanced topic at the bleeding edge of AI.
- Very engaging and fun projects to cap off each section. The AI nanodegree covered topics made popular in the media like game playing agents and deep dream images generated from neural networks.
- Pace and depth of the AI nanodegree was challenging enough to keep me interested but not too difficult to keep up with a full time day job.
- High quality material with mostly video instructions with well done animated illustrations.
- Very supportive and responsive team and mentor.
- Lifetime access to all material. Trust me you will want this.
I leaned on self learning for most of my coding career. AI warranted guided training because it's such a shift in technology and approach to solving hard problems.
Even with all my experience as coder, I struggled getting started in AI. Don't get me wrong, I'm sure given enough time I could have learned most of what I learned in the nanodegree on my own.
The key word is time. I have many ideas I want to start building with AI. I rather spend more time on building my next big thing than reading white papers. The Udacity AI nanodegree gave me enough AI in a short enough time.
I will still need to read white papers to keep up with latest and greatest. But now my reading will yield much more value and I'll be able to apply any new information faster.
Is getting an AI job your goal?
My goal was not to find a job in the AI field. I wanted to learn enough so I can start playing around with it in my own projects.
I did do a little bit of looking around the job market for AI jobs and found many good options. I found couple jobs that I feel I could qualify for given my new skills in AI.
Remember, advancements in AI are very new and industries are just starting to learn the power of what they can do with this new technology. Also, the expansion IOT in industry has create opportunity for AI solutions by providing the needed data to drive AI models. Business will soon need more than theorist and start seeking Ai practitioners who can implement Ai at scale.
I hope this shed some more light on the Udacity AI nanodegree and if it's a good option for you.
If you have any feedback or questions leave a comment. If you took the degree and had a different experience, feel free share also.
If you're a developer interested in getting started with AI or just curious about AI, checkout the mind-maps I put together. There's lots out there on AI and if you're like me and like visuals, you will love these maps. (click here)