I’ve probably been involved within the hiring process for data scientists a dozen times or more over my profession, while never being the hiring manager myself, and I actually have been closely involved in writing the job description for several of those. It form of looks as if this must be easy — you’re just attempting to persuade people to use on your job, so you may pick the one you want best, right?
Well, it’s actually more complicated than that. The general public on the market on the planet aren’t qualified for any given job, and even amongst those that are qualified, there could also be reasons they wouldn’t like working on this role. It’s not a one-way street; you don’t want just anybody to use, you would like the very best suited people, for whom this job would work, to use. So, how do you thread that needle? What must you write?
This column is barely my opinion and doesn’t represent the views of my employer. I actually have not been involved in writing any job descriptions my current employer has posted, for ML or anything.
To determine what to jot down, let’s break down what it’s a great job description is purported to do, for a DS/ML job or for another kind.
- Explain to candidates what the job is, and what they might do within the job
- Explain to candidates what qualifications you’re in search of in applicants
These are the bare essential functions, although there are several other things your job description posting must also do:
- Make your organization appear to be a pretty place to work for a various pool of qualified candidates
- Describe the compensation, work circumstances, and advantages, so candidates can determine whether to hassle applying
With this, we’re beginning to get into more subjective and complex components, in some ways.
In some spots, I’m going to offer advice for 2 different scenarios: first, for a small organization with few or zero existing DS/ML staff members, and second, for a medium or large sized organization with some DS/ML staff. These two might be quite different situations, with different needs and challenges in certain areas.
You might notice I’m using “DS/ML” loads in this text — I consider the recommendation here good for people hiring data scientists in addition to those hiring machine learning engineers, so I need to be inclusive where possible. Sorry it’s a bit clunky.
Firstly, for any organization, consider what form of role you’ve open. I’ve written up to now about different kinds of information scientist, and I’d strongly recommend having a look and seeing what archetypes your role suits into. Take into consideration how this person will fit into your organization, and be clear about that as you proceed.
The Small Organization
A challenge, especially for small organizations with limited or no existing DS/ML capability or expertise, is that you just don’t really know what your ML Engineer or Data Scientist may find yourself doing. You realize what general outcomes you’d like this person to provide, but you don’t know the way they’ll achieve them, because this isn’t your area of experience!
Nevertheless, you’ll still have to determine some solution to describe the role’s responsibilities anyway. I counsel being honest and up-front in regards to the level of information science sophistication at your organization, and explaining the outcomes you’re hoping to see. Candidates with enough experience and skill to provide help to will have the ability to conceptualize how they’d attack the issue, and within the interview process it’s best to ask them to do this. You must have some form of project or goal in mind for this person, otherwise why are you hiring in the primary place?
The Larger Organization
On this case, you have already got not less than a few DS/ML staff members, so you may hopefully call on those folks to let you know what the job is like day after day for an IC. Ask them! It’s surprising how often you’ll find HR or management not actually making the most of the expertise they have already got in house in situations like this.
Nevertheless, it’s best to also determine whether this recent hire goes to be doing mostly or entirely the identical thing as someone already in place, or whether or not they may find yourself filling a unique form of gap. In case your existing problem is just having not enough expert hands to do all of the work in your plate, then it’s probably reasonable to expect the brand new hire shall be filling a job just like what’s there. But, in the event you are hiring someone for a really specific skillset (say, a brand new NLP problem got here up and no person in your team knows that stuff thoroughly), then be sure you’re clear in your job posting in regards to the unique responsibilities this role may have to pioneer.
This brings us to a crucial point, as well — how much experience and which skills does your candidate have to have with a view to successfully do the job?
The Small Organization
- Experience: If this person is your first or second DS/ML hire, don’t hire someone without some substantial work experience. These folks will cost more, but in your situation, you wish someone who might be very self-directed and who has seen data science and machine learning practice done well in other skilled settings already. This might go without saying, but you’ve little or no in-house capability to coach this person on the job, so you wish them to have already got acquired training from other previous roles.
- Technical skills: But what skills do you really want to search for, then? What technical competencies, programming languages, etc will someone have to must pursue your goals effectively? Beyond ensuring that they will use Python, I’d recommend searching for advice from other practitioners in the sphere already in the event you can, to ask them what the skillset on your needs should appear to be. This changes loads, as it is a very fast-moving discipline, so I can’t let you know today what your Data Scientist or MLE might want to have the ability to do a yr from now. (I can let you know that asking for a Ph.D. is sort of definitely not the reply.)
When you do go in search of advice, be sure you’re consulting people who find themselves practicing DS/ML on the bottom, not only “thought leaders” or individuals who market themselves as recruiting whisperers. When you don’t know anybody directly who suits the bill, try searching through your network or reaching out to DS/ML skilled organizations. Take a take a look at other job postings that sound like what you wish, but be cautious, since these other postings might not be that good either.
Regardless, take this seriously — in the event you write unrealistic, unreasonable, or absurdly irrelevant/outdated skills in your job description, you’ll turn off qualified candidates because they’ll recognize “Oh, this company doesn’t know what they’re doing”, and that may defeat the entire point of this exercise.
An alternative choice is finding a freelancer in data science/machine learning to get you began, as a substitute of hiring someone yourself in any respect. There are plenty of fractional or freelance practitioners as of late, in addition to consulting firms that may take this whole problem off your plate. A fast google of “fractional data scientist” produces numerous options, but remember to do your due diligence.
The Larger Organization
- Experience: I’m an enormous believer in hiring less senior folks and training them up, in case your organization can handle it. Recent entrants to the sphere must learn one way or the other, and business experience is commonly the largest gap in a brand new data scientist’s skillset. Consider whether you really want to rent a Senior Staff Machine Learning Engineer, or whether you can promote internally and backfill a junior person. There’s no right or incorrect answer, but give it some thought as a substitute of jumping right to hiring probably the most senior level. We senior folks are each expensive and rare!
- Technical skills: As with the job responsibilities, that is time to ask your existing team for his or her advice. Don’t just ask them what tech they use, also ask them what they may wish to learn, if there was someone expert brought on who could share that knowledge. (These skills go in Optional or Nice to Have, not Requirements!) You have already got a DS/ML tech stack in place, in fact, so the brand new person might want to have the ability to work with that, but when there are adjoining or newer technologies that may profit your team, that is a great time to seek out out and potentially bring them on board. Don’t fall into the trap of asking for less than the identical stuff everyone in your org already uses, without giving any consideration or value to additional other competencies.
Also be mindful what your candidates have to have already got, in contrast with what they might learn on the job out of your team. Don’t inflate your requirements to make the role sound more prestigious, or to artificially weed out candidates, especially in the event you’re not paying commensurate with those inflated requirements, since you’ll be shooting yourself within the foot. You’ll be deterring the candidates who could be a great fit for the extent, and getting overqualified people within the pipeline who wouldn’t accept the pay you’ve available. And don’t ask for a Ph.D. if it’s not vital! (It’s almost never vital.)
It could appear insignificant, but when you’ve defined the role, picking a title to post really does send signals to candidates on the market deciding what to use for. I’ve talked in other pieces in regards to the evolution of titles in data science, and this continues to vary over time. But my shorthand advice, not less than today, is:
- Data Scientist: Not accountable for data engineering, pipelining, or doing their very own deployment, although they could be able to it. May do BI or analytics in addition to model development.
- Machine Learning Engineer: Chargeable for all or any of information engineering, pipelining, and doing their very own deployment. Does model development, but minimal or no BI or analytics work.
For leveling, I’d say this, as a really very rough rule of thumb, your mileage may vary significantly:
- Junior or Associate: Fresh out of faculty. No work experience. Possibly an internship.
- No Level: Could have had one or two skilled jobs or 2–3 years experience.
- Senior: 3+ years skilled experience.
Beyond that, there are higher levels that some orgs have and a few don’t:
- Staff: perhaps 6–10 years skilled experience.
- Principal, Senior Staff, etc: Greater than that. It varies so widely in several orgs it’s really hard to say.
So in the event you want someone who can do their very own pipelines, deployment, and modeling, and don’t need them to do analytics, and you would like them to have multiple years experience, then Senior Machine Learning Engineer is what it’s best to write. When you are in search of someone fresh out of faculty to do some modeling and analytics, but engineers can handle the deployment etc, you wish an Associate Data Scientist.
This recommendation is subject to vary as the sphere continues to evolve. When you actually need to jot down something special like Machine Learning Scientist, I’d advise against it unless you’ve a very good explanation as to why. Clarity and findability are key here — use the terms your candidates shall be aware of and looking for.
Now we are able to move on to your pitch: sell your organization as a great place to work, and share the compensation/advantages that you’ve to supply. We’ve spent plenty of time telling the candidates what they should bring, and what they must be prepared to do in the event that they get this job, but that’s not all that a job description is about. You truly also must be promoting your organization and department as an appealing place to work, with a view to get the very best candidates in your radar. This recommendation is generally generalizable to any organization size.
Don’t Lie
I actually have a couple of rules of thumb in the case of describing an organization to job candidates, in writing or in interviews. The major one is Don’t Lie. Don’t say you’ve a “fast paced culture” when it takes three weeks to deploy. Don’t say you “value work life balance” when nobody on this team has taken a vacation in a yr. And DON’T say “distant” while you mean “hybrid” for pete’s sake! You might think you’re just throwing in nice-sounding boilerplate, but these words mean something.
Feeling such as you got bait-and-switched into joining a corporation that may be a bad fit is awful. Consider it like selling a product — in the event you overpromise and underdeliver, perhaps you made that initial sale, but that customer goes to churn and be out the door with a foul taste of their mouth as soon as they realize their mistake. You then not only wouldn’t have that customer, you’ve someone on the market on the planet with a foul opinion of your organization who could also be telling their whole network about this experience.
When you can’t think of excellent selling points on your company that aren’t either lies or stretching the reality, then it’s essential to take a chilly hard take a look at your organization’s operations.
Being honest is not going to only make your eventual hire higher, but it is going to attract candidates who really do wish to work in an organization like yours. Everyone has different wants and desires from a job, and never everyone desires to work at a spot that “works hard and plays hard”. There’s not one right culture for corporations, and owning the culture your organization has will get you the candidates who might be pleased working there.
Value Diversity
One other essential key’s ensuring and displaying that your organization values and includes all of the angles of diversity amongst your staff. Your job description is the candidate’s first introduction to the way you’re taking good care of your people, no matter protected class or general diversity of experience, background, ability, etc. On this case, which means it’s essential to consider your selections of language very fastidiously. Unless you actually mean it, don’t ask for an “expert” in a skill set. Don’t say your candidates have to be “rock stars”. That is each deterring to candidates with reasonable humility about their skills, and likewise makes your organization sound, well, form of like jerks.
Note: the old saw that we’ve all heard one million times that “women don’t apply to a job unless they meet all the necessities” could be very, very drained, and problematic for a lot of reasons, nevertheless it does remind us to ask for skills you really need, not only a laundry list of wishes.
As a substitute, use inclusive language. I counsel writing your required qualifications in the shape of “Successful candidates can do ….” after which write motion oriented items like “construct machine learning models using Python” or “perform model evaluation using appropriate metrics resembling recall, precision, MAE, RMSE, etc”. Be clear, and make it easy for somebody to say “oh, I can try this” or “nope, I can’t try this”.
When you know your pool of potential candidates could be very homogeneous, for instance because not many individuals of color get college degrees in your field, consider whether it’s essential to take extra steps to get your job in front of those candidates. Take the time to post jobs on diversity-oriented job boards, and share your posting with skilled organizations for various kinds of individuals. In case your posting never gets seen by varied individuals, you won’t get varied candidates applying.
Compensation and Advantages
Now this could really go without saying, but be transparent and clear in regards to the advantages and compensation for the role. Give a compensation range even in the event you don’t must by law. When you’re not hiring in a state that requires a compensation range, you could think that isn’t a problem for you, nevertheless it actually is, because candidates with selections will prefer to use to postings where they will clearly see the pay is commensurate with their expectations. It makes you look exploitative to depart off a compensation range (or to offer a spread spanning $100k so the range is effectively useless). Get with the times and provides an affordable range.
Also, I already mentioned it nevertheless it bears repeating — be honest in regards to the working circumstances. Don’t advertise a job as “distant” only to disclose within the interviews that it’s 3 days every week on site. That’s also really bad practice and a rude waste of everyone’s time. Give candidates the main points they should make an informed decision about applying.
Beyond that, keep in mind that medical health insurance is essential to anyone in America in search of a job, and be as clear as you may be about what you’re offering. When you can list the insurance carrier, try this; it could help people know if their doctor or provider can be in network. It’s not an enormous deal to each candidate, but many candidates, including those with disabilities or health concerns (or dependents with concerns) will appreciate it.
Hiring for technical roles, including DS/ML, is difficult. This recommendation might all sound like plenty of work you’d slightly avoid, but consider: the choice is weeding through 1000’s of applications from terribly unqualified candidates, or candidates who would never accept the job. Do some work up front so that you’re not wasting your personal time and that of the applicants down the road. It’s not only more efficient, it’s also the moral selection. Applicants are real people and need to be treated as such.
To recap:
- Work out what the job would do (or what outcomes you must see)
- Work out what the experience level and technical skillset must be (not your dream wish list, but realistic needs)
- Write a job title that’s clear, accurate, and searchable
- Don’t lie about your organization or the job
- State the compensation range up front, and describe the advantages
Good luck on the market!
