The Algorithm Edit

Is it all just bean bags and table tennis? Are there white boards with maths everywhere? Do we all sit on Macs in a dark room creating mysterious robots? Some of those might be true, but I won’t reveal which..

I’ve been in the field about 6 years, which might not sound like a lot of time. But, I was here before the ChatGPT hype – so you know, that does class as ancient history now. I’ve moved around different positions, different companies, and even different types of AI research. I’ve picked up a few things along the way. On the whole, I really do love the work.

I love how many different skills you get to learn.

I love that it is fast paced, so you do have to stay on your toes to keep up.

I love creating things and I love to fix things I had broken earlier in the day. (Though I’ll admit it now, that I don’t love fixing someone else things they broke earlier in the day)

I even love the hype, that everyone gets excited by your work – and the generally everyone in the field wants AI to succeed in a positive way.

It’s a great area to work in if you are inquisitive and motivated. But it can also be stressful and extremely challenging. So today I’m going to share some of my experiences working in AI research – just in case you out there are interested in this field, or you are like me and want someone to relate to!

This role might be glamorous, well paid and just cool. And yes it can be *cough cough*, I have been to Silicon Valley. But some challenges might not be so obvious, and it’s not bean bags and table tennis all the time!

For those who work in AI, I would love to know whether your experiences align. For those thinking about working in AI, read on – I hope to help you be inspired by working in the field!

The Pace of Change Makes it Challenging, but Also Fun

Working in AI research is a lot of the time just really hard. Through battling the never-ending stream of new products and research breakthroughs, to maintaining skills in software engineering, scientific research and DevOps – I often feel like I just can’t keep up!

I go through phases. From endless days of paper reading – getting myself up to speed with all those recent breakthrough, but at the cost of falling behind in any project work.

Followed by weeks of focus time, building my AI application or running those experiments – this time at the cost of falling behind in the research forefront.

Balancing these mindsets is difficult, particularly for someone like myself who gets hyper-focused on one thing – I find it hard to take a step back.

Nevermind the fact that the theory and complicated ethics that comes with working AI are equally as difficult to navigate.

⚠️So a warning, I find myself in two conflicting mindsets – my project is going great, but I missed the newest paper. Or I’m way behind my deadlines, but at least “I know that the Transformers are no longer cool”.

⚠️ another warning, the job is actually technically very difficult. You won’t work for five years and find you know everything. You will never know everything, and will always need to learn a new technical concept.

Travel Can Be A Great Perk

But I can handle the constant rollercoaster of my mindset, because every now and then I get to travel and meet new people. Now not every AI research role will include travel, it depends on the company. But in all my roles, over my whole career, I have always been to at least one International conference a year – cool right!

I’ve been all over – San Francisco, Vancouver, London, Adelaide, Vienna… It goes on.

And it’s not just a perk of the job, nor a jolly, no it’s actually very important. I’ve found that going to these big conferences, hearing from the keynote speakers and the students presenting their new work – is extremely valuable.

The keynotes inspire me to carry on. The students have new ideas, which help me find new ideas. Meeting others, even just one connection at each conference has been extremely helpful and motivating.

So an awesome job if you love to travel and love maths.

💭 if you are looking for a role with lots of travel, this could certainly be one! Embrace the opportunities as they arise and throw yourselves in.

Research, Engineering or Analytics? The Landscape is Complex

A role in AI – what does that mean? There are hundreds of different types of AI jobs, and they can be quite different.

⚠️ so a warning, people who say they work in AI, don’t all do the same thing. The roles are different, their skill sets are different.

Research Creates New Ways

I’ve mostly working in AI research – which includes running experiments, creating new models and data processing techniques. Building janky prototypes, and showing customers what might be possible. Watching things just not work!

Research is where I like to be, I need to be able to write good and reproducible software. I need to understand the mathematical theory. I need to plan and run scientific experiments. And I need to be able to communicate this effectively with all sorts of people.

💭 I recommend going to research if you love to try things, you are maybe a little creative and you don’t care about having the shinest and best toy!

Analytics and Engineering is For Real People

Two areas that I’ve not had much experience. I’ve dabbled in product development – and it wasn’t for me! Turns out I am not a perfectionist and writing software tests and running proper MLDevOps is not my cup of tea.

But for those who love the engineering aspect, and building something that people might use – I’d consider ML engineer roles, or working in product development.

Data Analyst – I don’t really know what that entails! Its not research, its not engineering! Don’t listen to me here, we need to hear from an actual data analyst. All I know is they deal with real data, collected from real people – and perform analysis on that data to better understand their product or service.

💭 if you want to work with real people, for real challenges – then research might not be for you. Your work in research will take a few years and lots of collaboration to come to fruition. Instead, you might prefer an analytics or engineering role

Multidisciplinary For The Win

Your day will never be the same. You will never know everything you need to know in order to be the best researcher ever. You will need to be good at multiple different things to succeed.

Whats even worse, you’ll probably constantly feel like an imposter who has just got lucky. Or maybe that’s just me…

I keep pushing myself, even when I’m feeling overwhelmed – and it always passes, because the feeling of being an imposter is just part of the job I’m afraid. You work in AI, you work in unchartered territory, you might be walking in the dark – but so is everyone else.

⚠️ you will need to write software, know how to use a computer – and at a much lower level than just working Excel.

⚠️ you will need to be able to write reports, and give presentations – explaining to technical audiences and also to non technical people. This can be the most daunting task, but I also am so pleased after I have done it

⚠️ you will need to be mathematically minded, to be able to read papers with formulas in them, to write your own formulas.

⚠️ you’ll do really well if you can connect with new people, and form professional relationships – something I used to struggle with, but I am slowly getting much better at.

But, the best thing about needing multidisciplinary skillsets, is that your team is going to have such different skillsets. I love that I get to work with people who have had such different journeys to myself. They bring different opinions, that you have never thought of. Everyone will learn something, and its great fun.

So Thats the Truth

Thanks for listening to my rant today, I hope this was helpful for someone out there. Of course these are all my own personal opinions, others out there may have different experiences.

Its challenging, stressful and rewarding all in one – I’m here for the ride!

But its a fast-growing world, even if you are unsure there is no harm in trying – I’m not the person to discourage anyone from trying anything. You may love it as I do, you may hate it and never look back!

If you are considering joining the community – jump straight in! And let me know what it is that drew you to this world. You should also make sure to check out my Primer Series, here I will be discussing fundamental AI concepts – which you might find helpful when prepping for interviews!

For those in this community, what do you find challenging? And what do you find rewarding? I’d love to know and we should share with the others.

If you like to hear my take on things, you should check out my News, Opinions and Wild Thoughts collection.