Unlocking the Door

The Problem: Supply

The ongoing problem is simple to define. We need more developers who can proficiently build and maintain an ever-increasing number of applications, both internal and external facing. That’s it.

What makes this challenge a monumental task is that we already don’t have what we really need. So, we definitely can’t afford to take tons of our current developers’ time to train and mentor junior developers. The obvious solution – hire more senior developers.

And that is where reality comes crashing in like a hammer. First, there aren’t that many senior developers proficient in our technologies and frameworks in our area. Second, because the supply is so low, the salaries that this select group can command to hire on is substantial. Moreover, we don’t want the reputation of stealing every other companies’ best engineers.

One Solution: Build from Within

I can’t speak on behalf of my company – and I wouldn’t want to. But what I can speak to, is this one approach that I and my manager at the time, Joe Dunlap, helped design and deploy.

We agreed that we needed a way to get someone up to the point that they could be productive in the shortest period of time possible. Also, I need people to be far enough along in their understanding, that they will be capable of learning on their own once they get to a team – because there is no way I can teach everything that a given team does or uses and stay anywhere near my budget constraints.

Taking all of this into account, we looked back at some of the training that we both did in the military. The problem that the military faced when it came to manning was nearly identical. To counter these programs, they have built a myriad of different training programs. The programs were designed to be immersive, intense and often felt like weeks of ‘drinking from a firehose.’ But I had to admit something – they worked, consistently. In particular, the Navy Nuclear Program (which I graduated from in February of 2000) was built on this approach.

The solution that we came to was to do an internal software development boot camp. We selected Coding Dojo as a vendor to help us run it. I designed the curriculum based on the largest need that we had at the time, focused on front-end design in the React.js framework. For an end state target, I reached out to find out what the expectations were for a proficient React.js engineer at Facebook, as well as at other companies and built that as part of the end of course exam.

The boot camp itself was designed to be fast paced and intensive. It featured 2-4 hours of instruction preceded by daily coding challenges and followed by individual and group project work. Most candidates were consistently doing 60 hours per week (myself included). This pace continued for 7 weeks, beginning at fundamentals and culminating in delivering functional applications. The final exam itself was a handful of open-ended coding questions followed by attempting to build an entire small application in 3 hours.

How it Works

From a learning lens, the experiment was built on the idea of continuous immersion. If a participant spent almost every waking hour working on coding, thinking about coding, then there is a minimal information loss from day to day (which can be substantial, even from one day to the next as short term memory is wiped and only pieces of the day’s activity make it to long term retention).

This continuous immersion provides solid building blocks day to day that allow learner’s competency of material, concepts and methods to go from ‘unaware’ to ‘proficient’ exponentially faster than they would normally be capable of. This happens because they simply aren’t being given the time to forget much. Reducing the daily losses of new concepts and methods start out as small gains, but over a few weeks time these gains get to be very substantial.

Results

There were 18 people in our pilot attempt. Of these 18, 16 had demonstrated competence at or above our expectations. Of the two who didn’t, one was very close and one was not. We had a lot of learnings from the pilot overall and our net gains from the program were huge. People left the program excited about what they could now do – that two months ago seemed like an impossible ask.

Exiting the classroom phase of the program, most of the participants moved into development positions or development support positions. I helped arrange mentors for each of them, as well as an ongoing professional development plan to get each person up to the point of proficient engineer using their end of course exams as a baseline to build a personalized learning and development plan for each of them.

Overall, the pilot was a resounding success. We will be repeating it and I am excited to continue being a part of this amazing, transformative program that both helps people while helping the company at the same time.