Train Scientists to Become Automation Experts

Train scientists to program robots, code, and develop the advanced analytics we need. BioTech needs to train scientists to become automation experts.

Why are new skills needed in the lab of the future?

Labs are beginning to transition to automation dependency. Robots are performing increasingly complex manual tasks. Data generated by robotics is increasingly available immediately for additional experiments. The digital data generated from these labs are being directed to databases for storage and/or advanced analytics.

The role of laboratory scientists and engineers are changing.

• Larger numbers of experiments performed by automated systems at a constantly increasing throughput.
• Increasingly larger data sets that require databases to store, and record those connections.
• Advanced analytics implemented (i.e. machine learning) to process and model big data generated

The roles are changing because the workplace is changing. However, the workforce is not changing fast enough.  Automation experts are essential to keep up with the experimental needs of tomorrow. Coders able to write, adapt, and use applications that connect our robotics and data automation. A combination of both skills will help us rapidly assess and bridge technologies that change the way we work tomorrow.

Currently, only a tiny percentage of the workforce is capable of responding to those changes. This article describes many of those needed skills and why. There are three ways to increase skills in the workforce

1. contract talent
2. hire talent
3. train talent.

Contracting talent

Contracting talent from an outside vendor can be the fastest way to get new skills and tools working for R&D. Contractors are expensive, but, they can be a cost effective solution when skills are not needed long term. Here are things to be considered when contracting talent:

What is the true scope of contracted work?

• The hourly rate of contractors is high enough that hiring the talent out right will usually be less expensive. However, short duration projects are a great way to leverage vendor skills your workforce won’t needs for the long term.

What happens after the contractors finish the work?

• You need a plan to maintain or build on the completed contracted work.

How do you guarantee that the contractor is providing the right solutions for your lab’s needs?

• It is difficult to judge the value of work performed by a contractor with unique skillset. Consider how to assure that the contractor is delivering what we are paying for.

Hiring talent

Hiring scientists and engineers with automation and coding skills is the 2nd fastest path to acquiring those capabilities. However, there are multiple things to consider as it applies to hiring talent:

The grass is always greener… or is it?

• A wonder of hiring new talent is also a downfall. There may not be a path to assessing a new hire’s capabilities without some in-house talent to start with.

History matters…

• A pair of fresh eyes may make all the difference in addressing a challenge. However, those same fresh eyes are likely missing the context and history that shapes the problems and the required solutions. It takes historical context to be confident a new solution is the right solution.

What are the plans to replace this new skill set when a newly hired person moves on?

• There tends to be competition for people with rare skill sets. If you hired them away from a prior job, nothing is stopping them from moving on to green pastures. How do you stop from being completely at the market’s mercies?

Training talent: Train scientists to become automation experts

Training talent takes significantly longer than contracting or hiring it.  However, a training program can assure that there is a way to train more people.  There are a few things to consider as you train scientists to become automation experts:

• Is the training plan both timely and sustainable?
• Will newly trained staff use the skills immediately to assure mastery and retention of the skills?
• How do you assure that newly trained staff stay and create institutional knowledge for the future?

	Contracting	Hiring	Training
Time to utilize skills	Days→Weeks	Weeks→Months	Months→Years
Cost/FTE	Most Expensive/FTE but term limited	Neither most nor least expensive/FTE but long term costs	Least expensive/FTE but increased pay required to assure retention
Institutional Knowledge	None	Can be created if new hires train others	Created and maximized with creation and update of training programs
Project complexity	Complex	Varies depending on quality of new hire	Simple initially with increasing complexity
Major weakness	No institutional Knowledge	Amount of talent available	Amount to time to train
Major strength	Immediacy of skill utilization	Grow capabilities	Create institutional knowledge

Most laboratory automation work (robotics, database, and coding work) is performed by contractors. However, a tightening job market has made it more challenging to find contractors for reasonable hourly or project rates. It is also difficult to find and hire people with the prerequisite skills due to a tight job market. This means that the only option available is to train scientists to do that work. Retaining that trained workforce is the next big challenge.

Training Scientists: Active learning trumps passive every time

We created training programs for our liquid handling robots. It was realized that the training needed to be managed by our liquid handling robotic experts (Application Specialists). Two technology team were created consisting of the application experts of Hamilton STAR and the Tecan EVO liquid handling robots. They were tasked with the creation of a training protocol to train more application experts. The result was two training programs that were remarkable similar to each other in both scope and methodology. Each class consisted of 20 hours of class training that included hands on robot and simulator time. Students who completed class training then completed a mentored Journeyman project to become application specialists.

Making the active learning classroom

The most efficient way of learning a skillset or technique is to learn by doing. This means that students actively use the skills being learned to solve problems. Addressing actual problems from the workplace assures that the skillset being learned can be applied in the workplace. In example new trainees add and update existing labware definitions while learning how write scripts for different labware. This focus on active learning means it can take much longer to create relevant training programs.

It may also appear to take longer to learn in active learning classrooms. This is due to the reliance on trainees learning through trial and error, peer to peer learning and socratic training. However, students who complete active learning trainings immediately ready to initiate an independent journeyman project.

Journeyman Projects

Newly trained students become journeymen upon completion of their active classroom training.  Journeymen are capable of simple scripting and repairs of existing methods. They are not application experts until they complete a journeyman project.  The journeyman project demonstrates utilization of multiple items covered in the classroom training and may even establish mastery of methods and applications. New journeymen present their prospective projects and the completed projects to a team of application specialists. Journeymen are provided with mentors to provide feedback and support as they work on their projects.  We recently started asking that students identify a journeyman project before starting the classroom training. This can help students focus on topics covered during the active training that pertain to their project.

Active learning and journeyman projects

Active learning in a classroom setting is about acquiring and demonstrating a working knowledge of the basics.  Journeyman projects are about demonstrating mastery of a subset of skills to a mentor and other application specialists. The new application specialist is then considered capable to work scripts in production. They are also able to learn new skills without someone looking over their shoulder and mentor new journeyman.

Train scientists to become automation experts 

Will this training work to train coders and database scientists as well?

Almost all of our training work has been to teach scientists how to script for liquid handling robots. However, many of the skills learned include tools routinely included in coding. (i.e. defining variables, working with loops, and even database access and skills).

We are actively working to create coding training programs. We plan on updating this post with results from the first program this fall. If you are excited about this project, we would welcome hearing about your efforts to create a training program. Here are some considerations going into our efforts.

Standardize your training offerings

Our training programs for robot liquid handlers are possible because we standardized our systems. All of our Hamilton robots are running the same version of Venus. All our Tecans are running the same version of EVOware. A script written for one Hamilton can be readily transferred and modified for another Hamilton liquid handler without excessive rework.

This mantra needs to hold true for other trainings in order to make them self-sustaining. Select a small number of coding languages and standardize on them. The actual languages selected may well be driven by your organizations needs or existing capabilities. We have created this list that includes coding languages we thing are good standards. Fail to standardize, will make it challenging to create momentum and the institutional knowledge required to grow capabilities.

Dream big, but start small

It takes time and energy to transition an organization to recognize the value of train scientists to become automation experts. Think about the steps or stages required to achieve that transformation. Celebrate completion of each step or stage as a win. Without doubt, there is an event happening that requires scientists to take time to reorganize data to enable additional analysis. Let someone address that problem with a journeyman project. When they complete it, they will get credit for solving a problem and making other scientists’ lives easier. Multiple small wins can build into a transformational event. Including scientists in solving their own problems and thinking about coding solutions can change the lab.

Training and documenting needs to take priority early

Innovation in the lab of the future is going to be closely tied to the skills attained by the existing scientist workforce. There are many reasons for this, but the primary driver is that lack of other options to hire or contract these skills absent an incredible influx of capital or people. Even if there were a huge downturn in the tech sector, those professionals would need to learn the science or pair up with scientists to get any real work done. Training coders to do science will be at least as challenging as it is to train scientists to become automation experts, but we already have a large number of scientists who will need these skills in a changing work environment.

Basically this all boils down to the need to make training the top priority and documenting how that training is performed and produces results must occur. As the lab follows the currents of industry 4.0 the workplace will need to transform from a “continuously doing” model to a “continuously learning” model. This training and documenting paradigm lends itself well to that changing landscape.