What I look for in an AI employer?

My friends over at www.access-ai.com asked me these questions on what someone should be looking for when shopping for a job in AI/Data Science/Machine Learning employer.

Lately, as a manager of Machine Learning group within Deloitte I spend more time thinking about who to hire. These thoughts are my own so I will publish publically. With that said It is no secret we are hiring more data scientists than any other big-four firm https://www.linkedin.com/pulse/data-scientists-take-whos-snagging-most-talent-from-top-caitlin-crump

Before I answer, allow me to tee this up a bit. Many employers out there are looking for unicorns:

Thanks for the slide DataRobot http://datarobot.com

you should fit more or less within some of those areas. If not, stop reading this, it does not pertain to you.

The Chatbot Conference

What are the essential requirements that you have when looking for a job?

I have given a talk to several Universities (And even my high School) on how to transition from Academia to a job in Data Scientists in the last year.

Let’s talk about our capacities to handle change for a moment. Over 20 years ago when I transitioned from delivering Pizza to actually working on a printing press in college I noticed a huge shift in environments, expectations, and culture. Where you want to go depends often on where you come from. If I am straight out of a PhD program and I shift to working as glorified Data Sanitation Engineer that shift is often times epic and painful. Let’s take these point by point:

• environments: If I am working 90% behind closed doors before I transition, it may be difficult to go to 70% meetings and 30% behinds closed door. How collaborative are you? Do you like to wear jeans? Can you handle travel? Does the quality of the coffee really matter to you?
• expectations: Are you someone who needs a pat on the back? Seriously! Is constant affirmation important? How do you like your success measured? Are you a top down thinker who needs to be part of the vision process? Are you a bottom up thinking who likes to working off of observations and building hypothesis?
• culture: Do you like the people you work with now? Can you see yourself fitting in with the people at this new job? Beyond perhaps making some friends, do you see this as a place where you can communicate with others in the format you most prefer? Some places you need to schedule a meeting and have an agenda to talk with others, other places you can walk up to their desk and just talk. How accessible do you feel comfortable with making yourself? If someone calls you in the middle of the night is that ok?

Is there something which separates an average place to work from a great place to work?

Those who claim to have a great place to work usually say so because of their own personal expectations day to day being met with what the place can offer:

• Do you make the money you want?
• Do you get the affirmation you are looking for?
• Do you feel you are spending your time for a good cause?
• Do you feel others are contributing to your success with action?
• Do you like your desk?

These are 5 questions to ask. They look familiar? Yes, I stole them and repurposes them from 5 love languages. So the ultimate question is what in a job makes you feel loved. On the flip side, you may not love-to-work and instead work-to-love, so as always Work/Life balance is critical for some.

What office culture do you feel helps you to succeed?

Again, this depends on the person. I recall long-long ago when I moved from being a Printing Press operator to working in a the Pre-Press Department the manager handed me a ruler with my name on it. I didn’t realize the importance and symbolism till I later became a manager myself.

Certainly, folks want to be around others who give you the tools needed to be successful. If not, do you feel comfortable enough to ask for what you need? How comfortable are you working with other folks? Do you prefer to work with external parties (customer facing) or more internal? Again, how much time do you need locked away in your office while others shove pizza under the door?

Are there any questions that you will definitely ask of your employer when you’re interviewing for a role?

Whenever I am at a restaurant I almost always ask the waiter what is their favorite dish on the menu. They are a good person to ask as they are the ones seeing behind the scenes. I would consider asking:

• Where do you see company XYZ in 5 years?
• Do you like working here?
• What is the culture like here?
• What is your biggest challenge?

I do recommend asking at least one question during an interview. Do so that you show sincere interest vs you don’t want to interview them or put them on the spot.

Where do you look for jobs and employers? Which websites, agencies, events?

You oftentimes have to live with the method you used. For example, if I am online dating to find a life partner, there will always be challenge to get over that fact later on versus if the pair met organically. For some, organic-passive job search is the best. Others, like to have assistance from recruiters and job sites.

By the way, my club, Chicago Python User group, has a referral program http://www.chipy.org/pages/referrals/

There is a lot of exceptional talent working in AI today, and that becomes more the case as the field explodes. How do you distinguish yourself as a candidate?

Going back to the Unicorn diagram at the start of this article. How much of those three factors do you have:

• Hacking Skills: Do you know Python and/or R? Do you know full stack web development? Do you know how to use Git? Can you conduct Test Driven Development TDD? Are you good enough to implement scientific code in a production real time AI System? Do you know stuff like docker, lambda/events, dev-ops…
• Math & Stats: If I said my goal is RMLSE of .17 and an AUC for ROC of 93% with a 80% Precision and 70% Recall for 5 folds cross validation with a 20% hold out sample… would you know what the hell I am saying? hello? … hello? … Did you leave?
• Domain Knowledge: this one is huge when it matters. There are usually two different cross-sections in domains: industry or technology are. For instance, I may have domain knowledge in banking. OR, I may have domain knowledge in HMM (Hidden Markov Model). It is far easier to write Banking Models if I understand banking …duh!

When you go to market consider highlighting these three points for your potential employer. Don’t expect them to be highly convinced if you say you are a superstar in all three.

Is there an AI talent war, and has that made it easier or harder for you to find work?

We (Deloitte) have a whole practice behind data analytics behind things like attrition. I am not an expert in that area but contact me and I can help you find one. Although, one thing that has become clear is the movement of jobs seem to be different based on geography, size of company, business environment, and, of course, individuals happiness in that job.

For leadership (CTO, CIO, Chief Data Scientists) in this area like myself, the landscape has become very competitive. I get approached almost daily now through linkedin as well as even in person. Much to the point of this whole article, finding the talent, creating the star team, and keeping folks happy is not easy.

Maybe not war, but Wild-West a bit. The reality is that there are a lot of problems left to be solve. It has been in very recent years AI has actually become useful. I can’t imagine any sizeable company not going to battle to both get/keep talent and also start solving the world's problem with AI.

With all that said, now for those looking, go find the job of your dreams. Be prepared to adjust some to change. Do your best to bring best practices of AI to your organization. Do what you can to make the place you work a happy place for others. namaste!