The Three and a half Danny DeVito phases

Type into Google ‘hard work quotes’ and you’ll get pages of results all trying to achieve the same goal – to inspire you. They range from ones we’ve heard many times before, ‘Nothing good ever comes easy’ to those that are not as well-known but equally effective, ‘Those at the top of the mountain didn’t fall there’.  But it’s only when you’re forced to put these into practice that you realise it’s fricken hard to do the work to get there.

I took the decision many months ago to move away from I&A Strategy and rebuild my career in Data Science in the consulting domain. It wasn’t a decision I took lightly because it meant I had to put aside a role I worked so hard for, start again to rebuild my personal brand and area of expertise. But after many discussions, research and of course, a good old pros and cons list – I believed it was the right path for me.

To quote Kirk Borne (Principal Data Scientist at Booz Allen Hamilton) who I believe summarised it rather well:

“The excitement in data science is in the journey toward achieving three significant kinds of results: discovery, insights, and innovation.”

And if that wasn’t enough, the thought of one day being able to work anywhere in the world with just my laptop and a bit of WIFI was appealing.

 

My first project as a Data Scientist was in Security. Specifically, National Security. I joined the team and worked with two extremely talented Data Scientists and at the beginning I’m pretty sure I only picked up 12% of what they were saying (and that was mostly around ‘What shall we get for lunch?’). It felt like I had accidentally stepped into the cinema screen that was showing a foreign film.

But I stuck to it and didn’t let that deter me. I took scrupulous notes and studied in the evenings. My first assignment was to research Risk Score Algorithms and after a few days I craved something more challenging. I had a meeting with the Chief Data Scientist on a Friday to discuss working on a tangible deliverable. And he indulged me.

“Create an Object Detection Algorithm that uses Machine Learning”.

I looked at him. Unsure of what that sentence even meant.

“Any questions?”

I had about 1000. “Just a couple…for now.”

We discussed the task and before leaving the meeting he said,“I’m going to the US for a client workshop. Let’s have another meeting when I’m back next Friday. Have something ready for me then?”

“Sure” I said. I got thiiiis I thought.

SPOILER ALERT: I did not ‘got this’.

The week was chaotic to say the least. It was comprised of weekend working, internet searches that took me around in circles, and late nights in the office. At one point, whether it was the lack of sleep, the frustration of not making any progress or a combination of both – that I started to google, ‘Why are smart people on the internet so bad at explaining things?’

I needed a break.

I went downstairs, greeted the building’s security guard and went to the terrace. I was only two weeks into my first Data Science project, and I knew I was being hard on myself. But I also knew the nature of Consulting was to hit the ground running and become the expert in something without any lag. I went back upstairs to the office, packed my things and headed home. Reading articles on the tube about the subject – hoping something would suddenly click.

The next day came and as if by some stroke of luck, I came across a tutorial that made sense to me. It got the ball rolling and all the random things I came across in the last week slowly started to make sense.

It was enough to help me create an Image Classification Algorithm that used Machine Learning – which I presented back on the Friday.

“Good.” He said with a smile, “Very good. Now let’s really test your Machine Learning model”.

I held my breath as he played around with it. Hoping it wouldn’t break and let out a sigh of relief when my model provided the right prediction.

The next couple of months was spent adapting and optimising this algorithm for Object Detection, so that it could be productionised for client use. And yes, this again meant more weekend working and late nights in the office. But it was different this time, I knew it was possible even when things appeared impossible. And more importantly, I had a supportive team that were always willing to help me and provide guidance when I was lost.

It’s now that I would like to introduce you to “The Three and a half Danny DeVito phases” – what I went through in the last few months.

Choosing to do anything new can be challenging, we usually see the end result of someone’s hard work and not the journey they went through to get there. I think it’s important to share our experiences so that people have realistic expectations, so they don’t feel demotivated or worse, like a failure, if things don’t work or they struggle along the way. It’s normal and quite frankly I think we’ve all at one point gone through the Three and a half Danny DeVito phases.

So here are a few tips on how I managed the first hurdle.

1. Don’t be afraid to ask as many questions as you need or for feedback

They were in our shoes at one point, so they will understand where you’re coming from.

2. Don’t reinvent the wheel

There are rules, methods and even people’s experience that you can learn from and emulate. It’ll save a lot of time and put structure into what would have otherwise been a random approach.

3. Don’t Multi-task

When I was lost in the endless internet search loop, it was because I kept bouncing from one idea to another. (Granted, it was because each search result presented a plethora of new terms that I didn’t understand) but take it one step at a time. You don’t need to google every single word then and there.

 4. Break it down

I would suggest that you deconstruct the task into manageable chunks. For me, it made sense to first learn about the Mathematics of Machine Learning before working on the actual algorithm.

5. Share your findings

Write a report, give a presentation or set up a meeting with someone that could learn from your experience. Not only will you help them bypass the struggle you went through, it reinforces what you’ve learnt.

 …I know they will be eternally grateful (and not in a weird Jafar from Aladdin kind of way…more the endearing Aliens from Toy Story).

6. If you’re coming from consulting…

The good and the bad thing about Consulting is that you can be thrown into a role and expected to know the ins and outs of it almost immediately. Why is it good? After some time, you become confident in tackling the unknown and making it a success. But I found Data Science to be a little different, it needs time for you to build up the necessary skills. So, don’t worry if it takes longer than expected to make it work. You’ll get there.

I really hope this post has helped give someone the motivation to not give up and realise that they’re not alone. I’m still on the learning curve and every day presents new challenges, but if it’s something you really want to do, then keep working hard for it. And to leave you with another cliché, it will all be worth it in the end…

(Not sure what that bird is doing there or why someone left their laptop so close to the water…)

Share:

Leave a Reply

Your email address will not be published. Required fields are marked *