We’ve spent the past few years helping companies build their core data science functions by guiding them towards the best talent available in the market.  For many companies keen on gaining an edge in predictive technology, the dominant talent strategy has long been finding the best minds in this space and making sure the organization integrated this talent as deeply as possible.


But we have long predicted that once companies had mature data science organizations in place, they would begin seeking contractors to add flexible capacity to their teams.  In recent months, we have seen this prediction come to fruition with an uptick in clients requesting temporary talent for data science.


But here’s the problem: there aren’t enough talented data science contractors to go around.   For most people, full time employment is preferable for obvious reasons. And chances are that if you are a skilled data scientist in this job market, no matter what the circumstances are, you have a full-time, salaried position if you want one.  And even though some people prefer working independently and find benefits in the contractor lifestyle, we don’t expect the supply of contingent data science professionals to keep pace with the surging demand.




Dataspace maintains a network of skilled contractors ready to tackle your projects.  If you’re looking to add some firepower to your data science and engineering efforts in 2020, please let us know!  We can support requirements all across the United States and hold all candidates to the same stringent technical standards we apply to those for permanent roles.  All contractors are offered at competitive, all-inclusive hourly rates. Give us a call at 734-761-5962 or send us a message at info@dataspace.com!


How much should I pay for a data scientist?

Being that data science is a relatively new field, we often get questions regarding compensation for various data science related skillsets across different locations and industries.

Here are some example salary requests from qualified candidates we interviewed for healthcare related positions over the course of the last six months.  Despite being industry specific, we feel these salaries are broadly representative of compensation expectations across the data science domain.

We hope that by sharing this information we can bring some clarity to the hiring managers out there who aren’t sure what they need to budget when building out their data science organizations.  Stay tuned for more reports in the coming months!

Data_Science Salaries

We’ve built a business on applying thorough due diligence to each candidate we present to our clients.  If you are in need of contractors or employees for data science and engineering related purposes, let’s talk! You can reach us at 734-761-5962 x504.  Thanks for reading!

There’s a difference between a data scientist and a Python programmer who knows some data science frameworks


We’ve recently heard from a number of recruiters who’ve been burned. They were tasked with finding data scientists who cover the standard technical bases: Python (or R), Pandas, scikit-learn, and maybe even Tensorflow or Keras. What’s more, they found candidates with these skills. The problem was that in actuality their organizations didn’t need programmers—people who could write code. They needed data scientists—experts who could understand the business, identify where analytics could provide value, and could then write code.


So, there’s a role for both. Sometimes data scientists need supporting technical arms and legs (and brains) to bring their ideas to life (i.e., programmers). And, over time, these programmer types might grow into full fledged data scientists. Just don’t make the mistake of confusing one for the other. If you need someone who can interpret your business’s needs and then drive the coding to attack those needs, make sure your initial filters, and your subsequent interview questions, screen for those skill sets.


All of this points back to the importance of having a clear vision for new roles from the very start of the process – the more specifically you can define your goals for the position and the functions you hope candidates will be able to take on, the more likely you are to find what you are actually searching for. 



Struggling to find and screen data science and data engineering talent? If your current tools and vendors are working for you, congratulations! But, if you’re not getting where you need to go, if you’d like to work with a team of experienced experts who won’t waste your time, who know how to find and screen top talent, let’s talk!


We’ve got years of experience finding top analytic talent. In fact, in just the last month, we’ve placed a number of experts including both a director of data science in California and a manager of data engineering in Michigan. www.dataspace.com

We begin many blog posts by highlighting the acute shortage of data scientists and the resultant difficulty in attracting them to your company.  Chances are that if you are reading this, you are well aware of that fact and may have already found ways to make your organization more appealing to the data science candidates that are so sought after these days.  Maybe you have found ways to better market the cutting-edge problem solving your company is doing or refreshed your office space to compel candidates to consider your organization above all others.


All of these things will improve your chances, but a simple and oft-overlooked factor in the recruiting process can make an even bigger difference in how your organization stands out against your competitors: acting quickly when interviewing, hiring and onboarding data science talent.    


Too often, we see perfect matches between candidates and companies fall apart because the companies dilly dallied.  In the time that it takes for schedules to coincide, for hiring managers to respond to the availability request from HR, or waiting for someone to return from PTO, a range of things can happen to derail your interaction with a precious candidate.  For example:


  • They can overthink the prospect of working for your company and erroneously decide it is not for them
  • They can take the delays as signaling a bureaucratic culture and rigid management structure and think that it doesn’t reconcile with their own character or personality.
  • Most obviously, they can take a job at a company they may have not discovered if you didn’t take 9 business days to schedule their phone screen!


Long story short, there are dozens of ways for your recruitment efforts to collapse if you give them the opportunity to.  So what can you do to make your recruiting workflow more efficient? Here are a few things that can trim down the time from your initial interaction to a candidate signing the offer.


Limit the interview process to no more than three stages

A phone screen, phone interview and onsite interview should be the sole events that take place before making a determination about an offer.  Prospective employees want clarity concerning the process; adding unanticipated or excessive stages muddies the water and can create the impression that the company is unsure about what they want out of the candidate.     


Make HR a facilitator, not a gatekeeper

Internal HR and talent acquisition staff are extremely helpful when coordinating interactions between candidates and the people making the hiring decisions.  But it helps to allow direct communication so that HR doesn’t need to relay every question or concern that a candidate may have. Eliminating HR as an intermediary between candidates and hiring managers can reduce the time for concerns to be addressed, and also make the candidate feel included.  


Use video-conferencing instead of waiting for in-person availability

Instead of waiting for all of your traveling and offsite employees to be present in the office at the same time, leverage video calls when conversations can’t happen in person.  While virtual interactions don’t always convey the same amount of information, it is preferable to delaying the entire process.


Prioritize hot candidates over everything else

Communicating responsively with candidates sends the message that they are important to you.  People want to know they are wanted and valued and they will feel quite the opposite if they go days without hearing from you about next steps.


Looking to fill a challenging machine learning or data engineering role? Tired of seeing an endless line of crappy, unqualified candidates? Give us a call! We’ve got enormous experience in finding and screening top talent. In fact, our intensive screening process is the top reason our clients give for working with us. Let’s talk about how we can apply that same process to your needs!



As many a frustrated recruiter or hiring manager can tell you – a good data scientist is hard to find, and even harder to hire. Quality candidates frequently end up with multiple final round interviews (often within days of each other) leading to multiple offers. In such a competitive environment, your challenge is to figure out how to set your company apart and make it more appealing to top data science talent.  


The solution lies not just in crafting an appealing brand or corporate image, but has to encompass the more mundane steps of the application/interview process. A candidate’s experience during this critical phase can have a huge impact on whether they envision themselves having a future at your company. While certainly not all encompassing, the following points suggest a few key areas to focus on when you want to increase your success rate of guiding candidates from first impressions to a first day at work:


Communication: Prompt communication and follow-through cannot be stressed enough. This is not to say you must have detailed correspondence with every single person who applies. However, once you have identified a candidate who looks like they are worth your time, make sure they know you think they are worth your time! Lack of communication is not only likely to make a candidate feel disrespected, but it also gives the impression that your organization does not function efficiently, that ideas and people will get lost in the shuffle, and that it could take a long time to get things done. All of these negative impressions will increase the likelihood that a quality candidate will more rapidly engage with one of the other organizations vying for their attention.


Communication: Just kidding. But only sort of. Taking the principles of good communication one step further, we have the concept of accessibility. This element particularly applies for candidates who are in the second or third round interview stage. At this point they should be able to speak to and meet with the hiring manager (and potentially other members of the team), and they are going to want to see that this manager is engaged with and involved in the interview/hiring process. Managers are always going to be busy. However, finding the right next member of the team should be exciting! Managers who are too busy to get interviews scheduled promptly, and who can’t focus a decent amount of time on dialogue with candidates risk giving the impression that this will be the dynamic in the workplace as well – where leadership is not dedicated to setting clear expectations for their team (see next point) and is unavailable for questions.


Expectations – Know what the plan is for how this candidate will work with your existing team, and be prepared to talk about specifics. Candidates want to know what tools they will be working with and on what kinds of projects. They want to know what they will help your team accomplish in the next 6 months AND why these things are critical to the success of the enterprise. This kind of forward thinking helps them feel important and allows them picture themselves in the role. It also encourages them to start brainstorming about how they could contribute. They aren’t even hired yet, and they are already mentally engaged.


Offer Opportunity – not just a job. Quality data scientists are bright folks. They like what they do, find the technology they work with exciting, and enjoy learning about the use of new tools and new methodologies. They are hyper-aware that they are in a field that is constantly evolving, and they want to be able to keep up with everything that is new and interesting. A big deterrent is if they feel they will risk getting stuck in a box – in one specific niche doing one specific type of analysis as their skills stagnate. Data scientists will avoid environments like this because they know that ultimately, stagnation will result in the death of their career.


Facilities – Believe it or not, the physical work environment you offer a candidate can make a difference. The dream of many in technology is a career with the Amazons or Googles of the world. So, in recruiting, these places are your competition. While you may not be able to provide free gourmet lunches or views of the Pacific, you can provide an environment that says, “We’re about tech and we’re serious.” In an example of the importance of facilities and environment, we recently worked with a candidate who was facing two, competing opportunities from similarly-sized organizations in the financial services sector. One is housed in a 70’s-era building at the end of a shopping mall. The other recently invested in a beautiful, open-floor plan, technologically updated showpiece office. When making this investment management explicitly stated that it was intended to attract tech-savvy talent. At least in this case that investment paid off as the candidate, a super-strong data scientist, decided to take their offer, noting that the facilities were part of his decision.


Compensation – As much as we hate to bring it up, you know that comp is going to be important. Yes, people have different drives and some will value things like work-life balance more than others. Still, like it or not, someone who can make $160k / year is going to have a hard time accepting $105k just because of work-life balance. $140k? Maybe.


All this is not to say that you have to pander to candidates in order to get them to seriously consider your company. That isn’t doing you, or them, any favors either. But rather, the above points are meant to be an encouragement to step away from antiquated, assembly line hiring practices and instead approach the process with greater mindfulness and attention.  Quality candidates will return the favor.


At Dataspace we take pride in our ability to locate, attract, and screen top data science and big data talent. In fact, only 2% of all the candidates we see make it through our screen. So, if you’re looking to fill critical contract or permanent roles, give us a shout!


If you have experience in the tech staffing business, you may have noticed a recent rise in the amount of fraud and dishonesty in the marketplace. This is perhaps inevitable given the high level of demand for data engineering and data science skills. So, how can you separate those candidates who can walk the walk from the posers?


We here at Dataspace have come to realize that customers work with us partly because we intensively screen each candidate. In fact, our yield rate, the number of candidates who make it through our screens, is between 1% and 2%.


So we thought it would be helpful, and perhaps a touch entertaining, to list a few of the telltale warning signs we’ve seen and some of the steps we take to protect our clients from fraudulent candidates.  


Does it really make sense?

It is clear that there are candidates and staffing firms that just gather sentences and buzzwords from the web or from other resumes, and slap them together on the resume they send you. They assume that recruiters will get lost in the mass of text but see the right buzzwords (automated keyword scanning contributes to this problem) and advance the candidate.


The end result is generally a 4-5 page resume. But, if you dig a bit deeper, you’ll find that this text frequently just doesn’t make any sense. For example, you might find:


Competing technologies appearing on the same assignment – A recent applicant detailed a project built with Microsoft technologies on the Microsoft Azure cloud database service. But if you read closely you’d notice a single bullet point describing experience with Amazon’s Redshift. Now, it’s not impossible to have a project using both Microsoft and Amazon cloud databases, but is it likely?


We’ve seen the same when it comes to ETL tools. It’s possible for a company to use Informatica, Data Stage, and SSIS but, is it likely? And, is it likely that a consultant used them all on a six month assignment?


Statements about an assignment that make no sense for that industry – Earlier this year a candidate detailed his experience for a state government. One of the resume’s bullet points, however, detailed writing data movement jobs that handled multiple kinds of SKUs. Now, SKUs, or stockkeeping units, are a retail concept and not something that makes sense in most government environments. This was clearly a bullet point grabbed from somewhere else and thrown onto this resume.


We control for issues like these by doing a few things.  To start off, we are now very cautious about resumes that run longer than 2 pages. We don’t automatically reject them, but we do pay special attention to them. We also critically screen each line of a resume, looking for logical inconsistencies. And, finally, even if the candidate passes through those screens, we dig deep during the interview phases, forcing candidates to uncover which skill sets they really have and which terms were sprinkled onto the resume, like lemon on a five-day-old old piece of salmon to make it seem palatable.


Can the candidate answer very basic questions about where they worked?

We once interviewed a candidate whose then-current assignment was listed as “Walmart in Bentonville, AR.”  During the interview, when we asked him where he was calling from he said, “I am currently in Bentonville, Arizona” (AR is of course, the state abbreviation for Arkansas and Walmart is famously headquartered in Bentonville, Arkansas).  


In another case we asked a candidate about the street on which their recently-ended New York City assignment was located. Lo and behold, they couldn’t recall. We had to wonder, how one can commute to an assignment every day for 18 months and not know the street on which it was located? Interesting.


Asking simple questions about the weather or the commute can often be revealing and may give an indication that the candidate is fibbing about their assignment, or maybe even posing as a different candidate, which brings us to our next point.


Are you even interviewing the candidate who’s going to show up at your office?

Occasionally things will happen during the course of an interview which lead you to believe that the candidate is simply not who they say they are.  We’ve found that it’s commonplace in the staffing business for qualified candidates to participate in interviews for those that aren’t as experienced or polished.


We do all of our initial screening via Skype video calls. Why? because in the distant past we had a situation where the person that arrived at the client location was not the person we evaluated and contracted. This became instantly clear when the consultant who showed up had only the barest grasp of English and Business Objects, which was funny because, on the phone they spoke perfect English and answered advanced Business Objects questions.


We recommend that at least one stage of the interview process be conducted via video calls and that the call include someone who will be working directly with the contractor. Doing this will help ensure that the contractor who arrives on day one is the same one you interviewed.


Did the candidate actually work at the places listed on their resume?

Just last week a candidate had their most recent assignment listed as at a financial services firm and when we asked for a reference at said firm, the candidate hesitated.  After continuing to hesitate on three subsequent requests, we gave him an ultimatum at which point he provided us with a Gmail account for a contact who had no real association with the company. In addition, the candidate didn’t count on the fact that we’d done business with that firm and could ask our contacts if they knew him. In the end, despite his strong qualifications and interview performance, we had to pass.


Along these same lines, we have noticed that some contractors provide other contractors they know as references. So it is worth verifying that the person providing the reference is credibly associated with the company itself – make sure that the reference’s email address comes from a valid corporate domain and that there is a LinkedIn profile backing up the reference’s identity.




In the past month, a number of clients have mentioned that few vendors apply the level of scrutiny we document here to their evaluation process.  Many play a numbers game, simply finding and flipping resumes in their eager rush to fill roles as quickly as possible, neglecting to use even the smallest amount of professional care.


We’ve built a business on applying thorough due diligence to each candidate we present to our clients.  If you are in need of contractors, or employees, in the data science or broader big data and analytics space, and you’re sick of the subpar talent your current vendors provide (or if you just have some other, cool war stories), let’s talk! You can reach us at 734-761-5962 x504.  Thanks for reading!

The process of screening candidates for any role can be both daunting and time consuming, and even more so when the role in question is a highly technical position such as a Data Scientist. The ever growing list of tools and technologies that these professionals work with can read like a foreign language to recruiters who aren’t actively using these tools in their day to day work.


Obviously an essential step in the screening process is to note whether a candidate checks the boxes for the required skills and technologies that the employer is asking for. However, separating the great candidates from the mediocre requires going a step beyond a checklist. When looking to set apart a few candidates that stand out from the crowd and who merit a closer look, there are a few categories that I use as guiding principles:




I appreciate a quick, clear list or chart at the beginning of a resume that documents what tools, processes, and technological ecosystems a candidate has worked with. Obviously this is something I want to see backed up in the descriptions of work history… but if I have to read through paragraphs of text in order to discover whether or not a candidate has worked with Tensorflow, I’m likely to miss it.




Does the candidate tend to stay at least one year at each position? Or do they jump around every 6 months? Not to downplay the valuable service provided by short-term contractors, however a data scientist who has been at the same company for a year or more has had the chance to really get involved in their employer’s use of data science, take ownership of some projects, and see the bigger picture with how their work fits in with the company’s long-term goals. Longevity leads to greater breadth of vision, and valuable insight and experience that a candidate can bring to a new role.




No information dumping! No buzzword overloads!


A brief, clear explanation of work history shows that a candidate can truly wrap their head around the key points and overall mission of their experience. It shows that they have taken the time to sum up the core projects and processes of their work experience as opposed to just copying and pasting a laundry list from their current job description. The latter might be fine if an employer is looking for a short-term contractor to accomplish one or two very specific tasks. However, if the client is looking for someone to have the vision to help grow their program or  find creative solutions to their current problems, they will likely appreciate a candidate with a broader understanding of the scope of their work.


Being able to clearly and succinctly explain a concept or experience shows true mastery.


Attention to some of these “soft” or intuitive resume qualities is what an actual human recruiter brings to the table in contrast to a computer searching for a list of key terms… because in the end, a client is not looking to hire a checklist. They are looking to hire a team member.


contractors for data science

If you’re like many of the analytics leaders and HR professionals we talk to, you know how hard it is to build a strong data science team. Given that everyone is chasing a few fish in a small pool, it’s hard to find strong candidates with the skill-sets you need and to then win over the other companies fishing in that same pool.


However, many companies are focused on hiring permanent employees for their data science roles when, in some cases, contractors might provide a convenient, more readily-available solution while also providing other advantages.


Why contractors?


Project specific needs: Depending on the structure of your data science organization, contractors can help with project-specific needs.  You can ramp up the bandwidth of your team and then decrease it as your workload evolves.


Affordability: A common misconception is that contractors are vastly more expensive than full-time employees.  But once you consider taxes, benefits and bonuses, it turns out that the premiums associated with contractors are less than you might imagine.    


Larger talent pool: We are frequently inundated with resumes of people with some data science experience that returned to graduate school to formalize their skills and are now looking for their next job. Many of these people are immigrants and temporary residents. This group represents a large, relatively untapped talent pool.


While your company, like many these days, might eschew sponsoring immigrant visas, contracting provides you with access to this talent pool.  If you work with the right staffing agency, they can take on most of the paperwork and expense of sponsoring these people, eliminating that hurdle.


As we would for any position, we recommend a strong vetting process (or a vendor that implements such a process) to ensure that you’re getting what you need. But, with that process in place, considering contractors might greatly expand your options.


Capitalize on retention trends: Of course, most organizations are keen on owning the talent that they find and there is good reason for that – but hiring an employee outright does not guarantee that they will stay with your organization any longer than a contractor will.  Data scientists seldom spend more than 2 years with a given company anyway and you may well find that contractors are easier to keep with your company as they are not competing in the full-time employment market.


Try before you buy: As you’ve probably seen, interviews sometimes fail to reveal the true abilities or shortcomings of a candidate.  Contracting a person before bringing them on as a full-time employee allows for a more thorough evaluation of their technical aptitude as well as their ability to gel with the organizational culture.  


And while many companies are eager to hire data scientists, they don’t necessarily fully understand what those data scientists will be doing in their organization. Contracting can help you shape individual responsibilities of team members and overall data strategy without committing resources to full-time employees.  




With other analytic technologies, companies build a core of talent to interface with business users, set technology directions, and oversee delivery. Once this core is emplaced, these companies then fill out their team with contractors, allowing them to flex their team sizes to meet the challenges facing them at that moment.


Should data science be any different? Do you need a core staff of employees on your data science team? Almost certainly yes. However, we are seeing more and more organizations pursuing employee-only approaches when they would be able to move quicker and more effectively if they gave their staffing approach some thought and nuance.

Trends seen throughout the majority of 2017 persist in the current job market – low unemployment rates and rapid job growth across multiple sectors. These factors have combined to create a highly competitive hiring environment and a shortage of highly skilled professionals – especially in fields that have seen a higher than average job growth rate such as data science and the broader analytics space.


Making matters even more challenging, advanced tech professionals are now heavily sought after by almost every industry, not only the technical ones. Finance, insurance, automotive and manufacturing, marketing, food services… all of these industries now depend upon the work of data engineers and data scientists to keep abreast of a rapidly changing technological climate and maintain competitive advantage.As a result, employers face the daunting task of trying to recruit candidates with the requisite skill sets, all the while knowing that every candidate they interview is likely entertaining other job offers.


When you consider that even well-known companies located in prominent tech hubs are experiencing difficulty,  it becomes clear that companies in smaller towns or cities, or those lacking the budget to lure top-notch data scientists will lose talent to more appealing markets and industries.  


Given this environment, employers must rethink their hiring strategies in order to fill essential roles with qualified candidates and avoid falling behind.  Willingness to consider non-traditional candidates, such as those who are eligible to work with visa sponsorship can help bridge the gap and provide access to a broader talent pool – one that most companies neglect to explore.  An openness to sponsorship can provide a company with access to experienced, qualified candidates that other companies overlook due to the inconvenience involved.  


Not all visas are created equal


Keep in mind that there are several different categories of visa and work authorization documents in the United States and not all of them pose the same challenges.  In fact, some allow candidates to be hired immediately with little administrative trouble for the employer.


Yes, many employers are intimidated by the sponsorship process required for an H1 visa as there is a significant amount of expense, paperwork and uncertainty involved in it.  This leads them to reflexively screen out any and all applicants who are not citizens or green card holders. But companies that take this approach often overlook candidates that applied to the position on an Employment Authorization Document.  Not to be conflated with H1 visas, EADs are usually valid for 2 years, are renewable and are not employer-specific, therefore requiring no additional work on the part of the employer to sponsor.


Similarly,  F-1 (student) visa holders can apply to be authorized to work a period of OPT – Optional Practical Training. The initial OPT period lasts for one year, with an additional two-year extension often available for those who have earned degrees in certain fields such as science, engineering, technology or math.


Staying mindful of these differences expands the talent pool to include candidates with advanced degrees, experience and authorization to work without restriction for as many as three years before any additional sponsorship would be required of the employer.  


Flexibility is often the key to success and as demand for highly skilled tech professionals continues to grow, willingness to think outside of the citizen and permanent resident talent pool could provide not only an immediate advantage in filling essential positions, but an opportunity to invest in employees who could turn out to be lasting assets to the company.