How I got into AI, and is AI really democratized?

https://miro.medium.com/max/800/0*4EoZZSPk6tHOHXu3

Original Source Here

How I got into AI, and is AI really democratized?

As a career transitioner, I want to upskill so I stay relevant on the job market (and maybe land a job in tech?)

What’s a ‘job in tech’ anyway?

Back in 2020, I decided I wanted to change careers, and specifically — land a job in the lucrative tech world. Today I’m a step away from becoming a Microsoft Certified Azure AI Engineer Associate. And I want to share my rollercoaster journey with you, which could hopefully help you on your own transition trip.

No, my background didn’t do much to help me in regard to my AI learning journey. I have no formal training in computer science, neither a crush on STEM (Science, technology, engineering, and mathematics).

Briefly about me: I’ve been in the field of education for almost a decade, and I consider myself an education veteran. Essentially, I never left the field of education, I just swapped the roles, moving from the student’s desk to the teacher’s desk. I’ve been working on projects creating infrastructure for social rehabilitation and integration for children and young people with disabilities, and I’ve been teaching everything outside the STEM curriculum.

Frankly, being an engineer has never been my childhood dream. I have always been more into people rather than into technologies. It took me quite some time to realize that it’s their intersection point that makes technology brilliant.

Yet, my journey was not always sweet and easy. To elaborate, let me go a bit back in time.

Learning is a very messy place

When I decided I want to position myself in the field of technologies, this decision didn’t come with a clear and structured path on how to do that. Instead, my learning journey in its nicest view looked(s) like this:

I started off by playing around with simple data visualizations in Microsoft Power BI and a bit of Tableau. Data visualization is appealing for a beginner because it’s graphic, you see the results of your work immediately, they look nice and colorful and convey a clear message.

Yet, only 20% of the work is about tweaking good-looking visualizations. It turned out 80% is pure data science, including picking the right dataset, cleaning it, manipulating data, etc. And only when your dataset is neat and well organized, can you feed it into a data visualization software and start enjoying the sexy visual look that brought me to Power BI in the first place. About a month later I was able to produce beautiful, story-telling visualization dashboards.

But… Impact first, please

Alright, first milestone completed. At this stage, I realized I’m actually going after more interesting technologies in terms of business impact.

Throughout my work and learning experience, I came to the conclusion that women who go into the tech space are led by three key factors: Independence, Impact and Confidence. In short, we want to be independent, we want to create an impact with our work, and these two combined feed our confidence, making us thrive and strive for more.

That’s how I ended up with the decision that I don’t just want to be in the tech space. I want to be in the AI-tech space. Self-driving cars and humanoid-speaking robots sound so much more appealing.

Yeah, don’t judge me, we often start from this misconception. And thanks to the lingering around the AI space and communities, I actually learned that quite too many businesses go for AI solutions starting the conversation with the vendors as following: ‘Autonomous vehicles are great, AI is so exciting, let’s do AI’, while in reality they simply need a speech recognition solution, content moderation, recommendations solution or just a chat-bot.

Source: Unsplash, modified by me

I’ll share the story of my bumpy odyssey with how I got into AI in another article. (which would have been just a fiasco if I weren’t a combo of Buddha’s persistency and Dhalai Lhama’s consistency). In short, I dived into TensorFlow, almost sunk, then floated around for some time until eventually reached the shore of Microsoft Azure.

To make things clear early on: I love Azure for the democratization of technologies it brings. But since love for technologies does not imply reciprocity, I had to really get my hands dirty here.

Microsoft Azure, my savior

Having played around with Python and TensorFlow for some time, I finally understood that I can only move forward and progress through a really democratized tech solution, considering my rookie experience in the field. After my colleague Rafael Knuth pointed me to Azure, somehow the picture of the puzzle started getting less and less blurred. I only had to collect the pieces, and then start putting them together.

I immediately went ahead and signed up with an Azure subscription, got into the Azure ML Studio and my status went ‘Hold my cappuccino’. I found the user experience so very straightforward that it couldn’t go more beginner-friendly than that. (It did, it’s mid-2021 now, and Azure is everything but a static space).

There is the Azure AI Gallery, which most of the time, as a beginner, is all you need to try out and produce well-performing models. You just grab an experiment somebody else did, and since most of the time the authors of the experiments follow the industry’s best practices, you’ll get decent explanations on every step of the process. I tested with more than 20 experiments, even authored one: Heart Failure Deaths Prediction. (Authoring meaning it wasn’t already on the Gallery).

So far so good, but being able to copy-cat is not much of an intellectual and skills challenge. And I love AI for the intellectual leaps it makes and the transformative potential it bears. On top of that, simply copying experiments you don’t understand is not much of a success. That’s what brought me to the Azure AI Engineer Associate certification.

Understanding that AI is a big thing through reading the tons of articles claiming that, and realizing it after you’ve experimented with it, are completely different comprehension realms. There is this deceptive information bubble around AI, making us think we’re already living in 2090, especially if you look at it with my eyes at the beginning of my odyssey — or essentially, a bystander’s eyes.

I did a bunch of experiments asking people outside the tech bubble what AI is. And I got everything but realistic answers, and most had the impression that AI is reserved for those guys, the academia. And I thought to myself: it can’t be such a behind-closed-doors technology if I can already design and implement a simple end-to-end AI solution, from defining the business problem, ingesting the data and modeling to deploying the solution.

My point here is: Stay Hungry, Stay Foolish.

I’m all over these female-tech groups on social media and to quite a lot of women AI seems too distant and disheartening to dive in (and probably not only women?).

Indeed, most of the information about AI you’ll likely come across is eye-catching titles about breakthroughs and significant research achievements, but at the end of the day, the current level of applicable AI is business-friendly and practical. It’s something you can learn to do so your business can benefit from AI.

You don’t necessarily have to discover new planets and search for new drugs to do AI. You can simply start off by picking up some know-how, taking a smart learning path, doing some projects such as creating a chat-bot, a computer vision solution, a content moderator, etc, and eventually put that into practice in your job or in your business.

And to wrap it up in an even more encouraging light, Microsoft Azure really makes the AI kick-off a very neat journey.

There’s this over 110K-hits article on Hackernoon ‘TensorFlow is dead, long live TensorFlow!’ touting how TensorFlow 2.0 returns on a white horse bringing along the AI democratization. We’ve had long discussions with my colleagues on what AI democratization really is, and came to the conclusion that a democratized technology is Apple’s Siri, or Google Translate. In other words, a technology used from the dēmos, the people. Go ahead and try TensorFlow…

Jokes aside, I’ll be publishing more articles about my tech transition journey and I’ll be happy to share my experiences with you. Feel free to ping me on LinkedIn or shoot me an email at ava@knuthconcepts.com if you’re curious, or disagree, or anything.

Source: Pixabay

Jokes aside, I’ll be publishing more articles about my tech transition journey and I’ll be happy to share my experiences with you. Feel free to ping me on LinkedIn or shoot me an email at ava@knuthconcepts.com if you’re curious, or disagree, or anything.

AI/ML

Trending AI/ML Article Identified & Digested via Granola by Ramsey Elbasheer; a Machine-Driven RSS Bot

%d bloggers like this: