In the last year, I have seen a significant drop in entry level hiring. Listen to recent graduates and their stories reinforce what I am seeing happen across tech. Companies are not creating openings for entry level jobs like they used to.
This trend has been building for the last 5 years. What is going on? Three main factors are depressing demand for entry level talent.Automation of Entry Level WorkCompanies Playing CatchupIncreasing Project Complexity
As demand for recent graduates drops, I am beginning to wonder if it’s ever coming back. All three factors create an intense pressure to contribute at a high level even in jobs requiring little to no experience. It’s a setup to fail or burnout.
Talk to hiring managers and phrases like “hit the ground running” and “contribute from day 1” define what they are looking for. That mentality boxes out even those with internships or a year or 2 of experience.
Automation of Entry Level Work
In Data Science, much of the work usually given to entry level hires is either automated or distributed to new roles. Data wrangling is increasingly managed by the machine learning platform or third party tools.
Basic modeling was already a simplistic process using Python libraries. Now it’s gotten to the point where there is almost no code involved in EDA, regression, clustering, basic classification, and many other descriptive model development approaches.
Senior Data Scientists used to offload this to junior team members and review the findings once the groundwork was done. Now it’s a 10-20 minute process to get the results ourselves and there’s no reason to bring someone else into the process.
Over the last year, it has become more difficult for me to find parts of the project I can offload to junior members of the team. Most of the simple work items are not time consuming and doing them is not going to teach a new team member much about their job.
There used to be junior level work done supporting other teams. Much of the business uses advanced analytics but their implementations are usually simplistic. Low code and no code solutions allow for self service. External groups don’t wait for the Data Science team anymore because they have the tools and knowledge to manage many of their common use cases.
Unless the business has created a mentorship program to help junior team members be part of larger projects, there isn’t much work for them. I have heard some recent graduates talk about internships where they did not really get a lot of hands on project experience because the team didn’t have enough work they could do.
Companies Playing Catchup
Many businesses took a wait and see approach to Data Science. Now they are playing catchup with competitors who have spent the last 5 years building advantages using machine learning. Projects have aggressive timelines.
For most businesses, that means there is little time to bring new hires up to speed. Companies are afraid they are moving too slowly. They are not giving senior staff time to be mentors.
Senior leadership is paying attention to machine learning project delivery and how well these projects meet revenue goals. The pressure to produce is mounting on senior staff.
When teams hire, you can feel that pressure translate into who they are looking for. They want someone who can help now. Senior team members who are drowning in work cannot imagine finding time to teach even the most promising talent.
Pressure makes teams short sighted. It’s obvious the team needs to have a program to develop junior talent. When the deliverable schedule and revenue targets are all anyone talks about, it is difficult to get buy in for creating that program.
Increasing Project Complexity
The low hanging fruit and prototype phases of machine learning adoption have come and gone for most businesses. The project pipeline is filled with high complexity projects. RPA is handling the simple automation, so the Data Science team is handling more complex intelligent automation use cases. Requirements are complex and almost every model is highly customized.
The Data Analysis teams are handling advanced analytics and reporting, so the Data Science team is working on more advanced decision support systems. Business critical policy recommendations require significant evidentiary support. Model development is just the beginning. Decision support requires experimental design and management capabilities that few junior level hires possess.
Product facing models need to be reliable and built using software development best practices. They must scale and meet high availability requirements. They must integrate with existing product lines. Deploying and supporting models at scale is not junior level work.
Back to what I said earlier. It is really difficult to find ways to work junior team members into projects when nothing is simple. When every deliverable is product critical, it takes discipline and trust to delegate.
We Must Start Hiring Entry Level Talent Again
I am an advocate for hiring straight out of school for 2 main reasons:Entry level talent is easy to hire and comes at significant cost savings.Entry level talent is easy to train, and I am going to have to upskill or onboard in any case.
The time it takes to find experienced talent in the technical field makes entry level talent a comparable option as long as the business has a structured training and mentorship program. It works well to promote internally through the same programs.
However, my approach is rare. Why? Those training and mentorship programs are complex. Career progression needs to be mapped out. Senior team members must be given the time to mentor. More than that, they need to be taught how to mentor effectively.
Projects need space built in for junior team members. It takes work to carve out deliverables and pair junior and senior members up so they can work effectively together while still completing critical work items on time.
The program needs to be built out to set junior hires up for long term success instead of burnout from unrealistic expectations of performance or lack of support. It is a commitment of people and budget that few companies are willing to make.
That is our biggest challenge. If we do not continue to bring new people into technical roles, the talent gap will only get bigger. When companies feel like they are fighting for survival, it’s difficult to get them to take that long term view.