2 years ago, I changed careers and I couldn't be happier!

I moved from oil & gas to:

  1. Machine Learning Engineer
  2. Scientist for Machine Learning

I wanted to have a positive impact.
Now I work in weather.

It took a paycut, a lot of grit and a ton of luck but here's how:

What we're getting into today:

  1. Choose a good career
  2. Start Early
  3. Adjust the resume
  4. Adjust expectations
  5. Fill the skill gap
  6. Write applications, lots of them
  7. The first offer
  8. Finding jobs
  9. Interviewing

https://twitter.com/JesperDramsch/status/1554746359361818625

(Links are in the sidebar on the right!)

πŸ”€ Choose wisely!

There are careers that are easier and harder to get into.

Tech for example generally accepts more self-taught people.

In general, data professionals love a person that has worked with data.
All my applied science friends definitely have a leg up here.

⏲ Start early!

I focused on transferrable skills that would be useful in the tech industry.

Build specific examples where you honed these skills:

  • Coding πŸ ’ Build small apps
  • Data Analysis πŸ ’ Publish a nice PDF
  • Communication πŸ ’ Give a talk at a conference

I started publishing my codes on github
https://dramsch.net/github

I attended hackathons for projects
Quickseis

and wrote on kaggle

I gave talks
pythondeadlin.es

and made Youtube videos

Get creative! πŸ₯³

You can see all my projects from the last decades here:
https://dramsch.net/projects/

all my talks are here:
https://dramsch.net/speaker/

and media coverage is here:
https://dramsch.net/media/

πŸ“ Write a stellar resume

You have to translate your "niche CV" in something tech recruiters understand.

I believe applied scientists have a lot of data experience.
It's all about frame.

Consider hiring a professional to help.

This video helped a lot:


I really do think a lot of our skills are hidden in the framing.

Oil & Gas professionals talk about the oil field they worked in.
Worthless outside the industry!

Talk about skills & experience:

  • Data analysis on site
  • High-stress decision making
  • Ad-hoc coding solutions

Also, Ken Jee reviews data science resume on Youtube, so that's worth a peek:

Ken Jee Data Science Resume Playlist

I made a Skillshare class with some of these tricks.

This link should get you a free trial: Craft a Top-Tier Data Science ResumΓ© for Your Career Transition into Tech

πŸ”» Adjust expectations

We're leaving behind valuable things:

  • Our industry network
  • Niche expertise that pays highly
  • Skills in specific software and tools

Luckily they're still:

  • Valuable references
  • Background experience
  • Partially transferrable skills

I earned less in my job as a Machine Learning Engineer than I did as a PhD.

Granted, there were many factors, like a different country, pandemic, stingy company, etc.

But: I was prepared to take a paycut.

After 6 months in the new role job offers kept rolling in on Linkedin.

🧠 Fill the skill gap

I was missing some skills, like:

  • SQL
  • Business
  • Explainable AI

So I took courses to fill that skill gap.
There are great free courses on kaggle:
Kaggle Learn

And great free-ish courses on Coursera.

πŸ’Œ Write applications

I wrote 117 applications on Linkedin alone in 2020/2021.

Not to mention off Linkedin for the likes of Google, Deepmind, and Amazon.

Most I never heard from.

Some fizzled after first impressions.

Some went to last interview and reject.

It sucks.

πŸ’Έ First offer

I got an offer as a Machine Learning Engineer.

I was looking for Β£60,000. National average without London is 55. They low-balled 40 and adjusted to 45.
I accepted.

Why?

I knew I'd keep searching.
Also, the line item of "someone in industry gave me a job" on the CV helps.

Don't accept everything, obviously.

The team and my direct manager were amazing.

In that job, in 6 months, I worked/helped on machine learning with:

  • GPS jamming
  • Satellite imaging
  • New satellite sensors
  • Drone navigation and vision
  • Point-cloud data from trains

After about 6 months in that job I saw a shift.

My applications got beyond the CV stage.

First calls. Interviews. Take-home tests.

Still exhausting, but at least it wasn't complete silence. πŸ€·β€β™€οΈ

It has been lovely not writing an application for over a year.

Funniest thing.

I still got rejections months after starting the new job.

The last one was from Twitter, I think.

It's really a lottery and you're trying to get more tickets, but here are some things that helped me.

πŸ”Ž Finding jobs

There are obvious places like Linkedin and Indeed.
They're terrible, but it works for getting job alerts.

I found ianozswald's newsletter with jobs very useful.
Got many interviews from the "not-entirely cold" jobs there:

notanumber.email

It's always easier getting a "warm lead".

This can be:

  • Ian's Newsletter
  • Recruitment Events
  • Outreach opportunities
  • Tech events like Google I/O or PyData

It will always be easier if you're not just a CV in the stack.

Don't ignore the "small jobs".

Truth be told, I didn't know what the ECMWF was until I interviewed. 😳

I got really lucky (as so often).

The interview process wasn't "tech-y", and they appreciated my prior expertise in physics and communication skills.

Now I'm here!

🎀 Interviewing

I'll say it outright:

Tech interviews are stupid.

Deliveroo rejected me after 7 interviews and a take-home
Amazon rejected me after 2 calls and a final full day of interviews Γ  1h each
lyft rejected me after 2 calls and an interview day

They say tech interviews are a skill.

Get used to rejection and a lot of "wasted work".

Things I learned:

  • Learn Whiteboard coding interviews (leetcode etc)
  • Ask questions; they expect a conversation
  • Be prepared for insignificant details
  • Give the simplest answer

Tech interviews don't have space for nuance.

I was asked if you can use Mean Squared Error for classification tasks.
Technically you can, when the classes are ordinal, and you define bins on the regression output.

"Yes"

That's the wrong answer.

Don't make my mistake.
Stay Simple

🏁 Wrapping up

Changing careers is difficult.
But it's worth it.

Key takeaways:

  • Fill in Skill gaps
  • Find warm leads
  • Translate your CV
  • Get used to rejection
  • First offer is the hardest
  • Interviewing is a skill and sucks