I’ve been increasingly aware of an inclusivity in tech problem. In 2017, I came across a non-profit in LA that was tackling a big, complicated part of the problem: career opportunities for young people of color in low-income communities. Their program met kids in middle and high school and taught them how to code. Alongside that, the kids learned the soft skills they’d need to be successful working in tech. The non-profit would make sure they had laptops and WiFi, make sure they showed up to class, and otherwise open the doors to a career in tech. Paid internships hadn’t yet crossed my mind.
A non-profit teaches kids to code and sparks their interest in computer science, which will eventually expose them to high-paying jobs. Sounds good, right?
The Problem
What I learned next gave me pause. Teaching the kids to code was working — not overnight, and not at the scale we need, but kids were successfully completing the program. However, the graduates were hitting the next layer in a deep, generational quagmire — without middle and upper-class onramps into tech like personal networks and internships, it was really hard for these kids (now young adults) to get their first jobs.
Even if they could get internships, many of these students were responsible for contributing to their household incomes as early as possible. Working an unpaid internship to get a better job later was simply untenable. They needed a way to continue building their skills, and they needed a way to get paid doing it. If there’s one thing I remember clearly about learning to code, it’s that it took hours and hours of generally “unproductive” time that wouldn’t pay off for years. This could well be one of the (many) underlying factors to the inclusivity in tech problem.
I couldn’t stop thinking about it. If you can code a little bit, what are the limiting factors preventing you from making some money with that skill? I’ve used services like Upwork and Fiverr that let you connect with developers from all over the world. You definitely get what you pay for, but the general quality level is not as high as I’d like, nor is the transactional nature. I also don’t love the security and privacy implications of giving strangers across the globe access to my codebase, or worse, my credentials. And let’s face it, I don’t really want to be someone’s first client. By definition, the chances of an amateurish outcome are pretty high.
Building Modularly
What I wondered was this: can you pay these early-career developers to build standardized chunks of code? The benefit of writing in these discrete pieces is that the developer doesn’t have to build an entire app; they can tackle writing one small piece in a larger working app, so you’ve got an appropriate amount of responsibility for the coder’s skill level.
At the same time, the coder can get a sense of completeness from having responsibility over a specific area. And if, as a platform, we regulate things like the way we handle passwords and sensitive data, these app ingredients, or “modules,” would be relatively safe to use.
By building modularly, we can also combine the pieces in spontaneous new ways; a user interacting with these modules in a no-code app could build something quickly, securely, and fairly customized to their needs. With as few as 100 modules, there would be as many ways to combine them as there are atoms in the universe. It would be easily 10x faster than building from scratch, and with fewer bugs because there’s less new code being released per change. And if all that were true, it would actually be pretty fun to use.
No Code, with Source Code
I’m generally leery of no-code, but I decided it might be acceptable if we gave users access to their source code. The platform would be incentivized to keep users by being great, instead of being locked in against their will. And if you hit the limitations of the platform or think you can run it cheaper/safer yourself, you just eject a NodeJS app. Exposing the source code of the apps we produce also supports version control using modern amenities like GitHub. Most importantly, I wanted the code for the apps to be written the same way I would write it myself. Just faster.
The Paid Internship Experiment
I was finishing a mobile game as a side project. While doing that, I wanted to see what it was like working with graduates fresh out of these learn-to-code programs. I was a little scared to A) work with new developers of unknown skills, and B) see whether it might be awkward that we were coming from such different backgrounds. Pushing myself through that bit of hesitation, I reached out to a friend of the non-profit and asked — are there any recent graduates who could use a paid internship? I got two applicants. We did video interviews. I made up a JavaScript assessment so I could see what they knew and watch how they tried to solve problems. I decided to hire them both for a two-month, part-time stint, paid at $25/hour.
For the internship itself, they worked 20 hours per week. I gave them tasks and started coaching them, employing a mix of independent time and “office hours” during which I’d be available in real time. They fixed bugs & designed levels for the game I was building. We designed experiments and reviewed analytics to see lean startup concepts in practice.
The Results of the Paid Internship
They needed a lot of coaching. I’m at my best when I have someone to show up for though, so it worked out. I tried to balance keeping the project moving while going deeper on the skills they’d already been learning: how to vet ideas, debug an app, manage a repo, and work on a team. They were twenty-somethings that wanted to get better at writing code and get a job in tech; I’m more than a decade ahead on that same path. We had plenty in common.
At the end of the two months, one of the interns parted ways. The other took me up on an offer to be a co-founder at my next company. The game we’d worked on was finished for the time being, and I wanted to start writing this no-code platform I had in mind. I also wanted to take the “paid interns building modules” concept another step further. We spent a year exploring different ideas, building prototypes, and trying to build a UX based on modules. Something that felt enough like real code to be powerful, and different enough that it would feel approachable to the average office worker. Inclusivity in tech doesn’t just mean people from different backgrounds, but also skill levels.
False Start
At one point, it felt like we were on the verge of product-market fit, and we picked up some angel investment. We added a co-founding marketing officer and a co-founding engineer. A board of advisors. We ran the paid internship program again with two more people; they built some new modules for the platform and were a pleasant temporary addition to our team.
The original question we had — could coders at this skill level build these modules? — came back with a resounding, “Yeah, pretty much!” Someone mentioned our paid internship program in a career panel for women in tech and we got eight applications in a weekend. Whatever it was we were building, there seemed to be an appetite for the paid internship program. This seemed like a step towards improving inclusivity in tech.
No Product, but a Little Hope for Inclusivity in Tech
The paid internship program had tentative, anecdotal legs to improving inclusivity in tech. However, despite everything we tried on the marketing front to get people interested in the no-code platform, nothing really took. We were really struggling to find a community of users to tap into. We got close to landing a few big clients, but they all petered out. One to a vertically oriented competitor, the others because of how early we were. SOC 2 compliance is tough for a four-month-old, five-person team with no revenue. We had to scale the team back down while we continued the quest for product-market fit.
Around that time, we got one gratifying piece of feedback — I got an email out of the blue from SpaceX. One of the original participants of the paid internship was applying for a Software Engineer role and listed me as a reference. I talked to the recruiter and gave an honest assessment of their strengths and weaknesses; I probably gushed a little, but I really tried to hold back how much I wanted it to work. The last thing I wanted was to oversell their skill and let them end up over their head. Drumroll… They got the job!
Paid Internships as part of the onramp
Whatever small role we played in their journey, we got to be part of the ramp getting this kid (well, they were a kid when they started the original learn-to-code program) from east LA into a professional role in tech. Would they have gotten the job at SpaceX without our paid internship? I have no way to know, but I’d like to think it was a notch in their favor.
We’re still trying to figure out the right use-cases for the no-code platform. Building in modules seems to work. Building with paid interns from underrepresented backgrounds seems to work. If we can unlock that at scale, it’s a benefit to society. A step towards true inclusivity in tech. And we’ll have goodwill with the next generation of a scarce resource: great engineers fluent in our way of thinking. That feels like something worth tackling.
We’ve basically stopped marketing while we try to figure out what we’re building. How we position the platform and how it monetizes are currently an open question; we haven’t figured out exactly who our user is and what problems they need solved. We think it might be in the automation space, streamlining business processes.
Going Forward
To continue exploring this path, we need two things:
Real-world problems to solve. What processes do you wish were better at work? Even if you’re not sure how to solve them, hearing about the pain and frustration is helpful to us. We’ve heard about onboarding team members, syncing inventory levels, creating purchase orders, and ingesting data. What else?
Paid intern applications. We’ve got a few slots open for the next cycle of the paid internship program. Our next cycle will have four people, now that we’ve seen two work. We’d love to meet people from underrepresented backgrounds who can write a little code and are looking for their first job in tech.
