Robert Sanders · Self-taught dev turned founder
From getting a job with no CS degree to building and managing developer teams
Robert Sanders
Bristol, UK
Studied linguistics at uni
Published his first iOS app just by following YouTube tutorials
Co-founded a startup that raised around £500k
Paid himself next to nothing for 5 years as the CEO
Runs a different flavor of Linux every week
Behind the scenes
I consider Rob (LI: @robert-sanders) one of my best friends. We met in 2016, during a master's course/business incubator, and ended up living together for four years.
We went from not knowing each other, to having dinner together every single night due to our synchronized cooking schedules. We used to eat and talk about coding, life, the universe, and everything. We never reached a conclusion, so maybe the answer is really just 42.
So Rob, what's your backstory?
I tried to learn how to code at school when I was 12 using Visual Basic or some precursor to that. No one knew how to do it, except one teacher who tried to teach me but I don't think he wanted to, so he ran away.
Then at uni, I googled “How do I write code“, thinking there's one language that you just learn and then you write code. So I got into this whole rabbit hole of which languages do what, which one you should learn first, all this kind of stuff. And I wrote a little program in C, a little text-based game.
I was just really into it. I was hooked.
What drew you to coding?
I was always interested in how stuff worked. So anything really, but computers seemed to be the biggest thing. Because there's so much stuff going on. I was like, how on earth is this working? And I just naturally found that fascinating.
When I got into actually writing code, hitting “run“ and then seeing it work and do stuff just seemed to be so cool. So I just got addicted really quickly. And then all I wanted to do was write more code and do stuff.
Do you have a CS degree?
No. I studied linguistics. There's a surprising amount of overlap between linguistics and code, but it's not really code.
There's this thing called the logic of language, where you break down English or any other language into logic. And that's very similar to how you deal with computers as well. Like “if this, then that“, that kind of stuff.
When did you become a developer?
Throughout uni, I created this program to help my parents who were farming collect data. When you inject a sheep with some medicine, you have to track it, all that kind of stuff, and all the programs were really crap. Or they didn't like them anyway.
I was making it really easy, right? Because they weren't techies, they had big fingers, they wanted big buttons, all that kind of stuff. So I just made that using Objective-C, before Swift was around. And then after uni, I published it and actually started a little mini company.
But I was young, I didn't know what I was doing. So I made some leaflets and I went to the local markets and things, and I'd show it to farmers. But the farmers would be telling me how they were really proud that they couldn't even write let alone use a computer.
I think after about a year, I realized that:
I didn't know enough about programming and about running a company, and
these farmers weren't going to pay me
No one taught you how to code. How did you make an app and publish it?
I did it all through YouTube.
There's this awesome YouTube channel called AppleProgramming. And the guy did C lessons and then Objective-C. And that was really cool. But that was everything. Basically, I learned everything through that.
Inversion of control, dependency injection, security best practices, microservices, all that stuff came way later. At this point, I was throwing everything in a file and making it work. If it's got a bug, I'll find it because it will crash when someone's using it.
How did you get your first job? What was the salary?
My parents were trying to persuade me to get a job in a management consultancy like PwC or Deloitte. And I was like, hey, give me three months to try and find a job as a programmer. And then if I can't find a job as a programmer, I'll go through the other process.
I remember I sent out 50 or more applications. I got two interviews, and one job offer.
At the time, I had the app that I'd made. So I remember going into the interview where I got the job. The guy had downloaded my app and was playing with it, so he could see that I could write code.
He knew that he was hiring someone who at least knew how to make an app, right? So that helped a lot.
I know I got a pay rise to £25k. And I think it was £22k that I started on.
Did you learn a lot there?
Yeah, I did. I stayed there for just over a year. I learned more afterward, but it was a really good starting point. I really enjoyed what I was making.
I used probably four or five different languages there. Probably about three different technologies and SDKs. It was basically like a playground. It was good fun.
How did you get your next job?
I went looking for a job whilst I was still working at the last place. But it was through LinkedIn, all LinkedIn. I think I uploaded my CV to cv-library or something. And then a recruiter connected with me on LinkedIn because he saw my CV up there.
He said, “Hey, I've got a good job for you in London at Capital One, which is a credit card company“. And it was quite good. It was a £10k pay rise. And it was going to be in London with all this cool stuff. I think I had a weekend, that was it, between jobs. And then I started my new one straight away.
Was there a technical test?
There was an online test, and then I had to go into the office in London for half a day. I did like this pair programming thing. It was quite easy.
I found that the first couple of jobs, they were looking at my degree and stuff. But as soon as I had that first year of experience, no one bothered.
It was just like, yeah, cool, you can code, it's fine.
What did your day-to-day look like?
I was a level two iOS engineer using mostly Objective-C with a little bit of Swift as well. But this was big company stuff. They had some flexibility, but this is a finance company. They're tightly controlled and everything was heavily tested.
The first thing I ever did was change a tiny bit of text in one of the buttons. I changed it to “Press Me“ from ”Click Me” or something like that. And that was like a big thing. I had to run through all the different processes and passable tests and all that kind of stuff.
And like every test you can think of, they had it, and they had it in all these pipelines. So I learned a lot. That was really good from a process point of view because you learn all the standards, the best practices on testing, and AWS security and stuff.
But if it takes them six months to implement a tiny feature, I just can't work like that. So eventually I quit.
How did you go from that to starting your own company?
I was at a point where I was like, I don't know what to do next. So I was doing a little kind of pros and cons thing. One of them was to become a farmer. One of them was to go and work in Barcelona. One was to stay in London and work for a different company because there were some pay raises available. And the other one was to go and start my own company.
Then I saw this master's program down in Falmouth willing to pay people to start a company. I was like, well, actually, this is a possibility now. I can do this now. Thinking I could, like previously I thought, I had no chance to do it, but now I was like, yeah, this is good.
All my mates in London were like, what on earth are you doing? You're already successful.
Which is quite interesting for them to say. I was like, I don't feel it, I feel like I'm bored. This isn't success to me. This is boring.
So, long story short, I ended up starting a company with one more founder. We managed to raise cash, we managed to get clients and we earned money. I worked out the other day, in total, we raised about half a million, and we also earned about half a million.
But we were always kind of fighting a couple of things that just weren't quite right. I felt like it was an uphill battle.
But it was good. Like overall, it was a really good experience and I wouldn't have been able to do that in the first place if it weren't for the farmers and the incubator.
The farmers?
Yeah. The fundamental purpose of Glas Data (our company) was to bring in any different data related to agriculture, regardless of how it looked, and then analyze it and say, if you want to increase your yield in this area, you need to adjust these variables to achieve that. If you want to decrease your carbon footprint, you need to adjust these variables.
Like a massive neural network, basically. Like a massive predictive model. But we never got that far sadly. We were getting close, but we didn't manage to build that bit.
Why do you think Glas Data failed?
Fundamentally, I think the product wasn't put together right, and we didn't sell it right. It was too complicated and too big, and no one wanted it as a product. There were parts of the product that sold well, which people loved and I think could have made money and been successful.
But it was too big, it was too generic, it was too complicated, and no one wanted the entire thing.
If I had to do it again, I would split out what we built into maybe 30, 40, or even 50 different products. I would choose one and I would build it and I would get money in on that and I would make that a success. And then, I'd go to the next one and the next one.
And of course, there were other things that went wrong, right? I was unlucky with some of the people I ended up working with. I was unlucky in some of the people I tried to do business with. And also farmers. Farmers don't want to pay. So I've chosen a really hard bunch of people to try and get money out of.
Did your role change from developer to manager?
Yeah, well, sort of. I think even at the end, I was writing a lot of code. So really, the company was my co-founder selling stuff with a couple of people helping and me managing a team of devs. We had three other developers and myself.
Occasionally I'd be rolled out to talk to investors or like people presented to buy it. But most of my time was just managing code tasks and then coding myself. I didn't stop coding.
How do you manage people effectively?
I'd say take the time to understand the people. You're going to have some guys that are going to go really quick and be really energetic and some guys are going to be really introverted and be really slow and like want loads of detail. Some guys will get annoyed if you give them too much information, and some guys won't do anything until you give them loads of information.
I don't think you can have teams that are too big either, because you need to be really flexible in how you work with people, otherwise, they don't work very well.
Would you do something differently if you had to manage a team of developers again?
I think I'd do a little bit more process. I think I'd do a little bit more complexity reviewing and a little bit more automatic checking and stuff. Not too much because it can bog you down and get annoying.
But apart from that, quite honestly, I think one of the best things that we had at Glas was the dev team and the code. Like the product failed for many reasons, right? But the code was good and the dev team was good. I'm pretty pleased with how it went, to be honest.
Since £1 Million went through the company, were you able to pay yourself a decent salary?
No, no, I didn’t do that. So I started off with no salary. Then I drew like a thousand pounds a month, then £1,200. And then I paid myself a salary, I think it was £25k. And I paid myself £25k all the way through to the end.
So you went from earning £50k at Capital One, to no salary, and then eventually to £25k?
Yeah. [Rob laughs] Yeah, quite depressing isn’t it?
But I mean, so Glas went down. But then I had five years of CTO on my CV. So I then went and became a CTO in another company and earned good money. So it paid off in the end. And the experience was amazing.
But yeah, it was five years of pain, basically.
My investors called it road rash. They wanted to have a CEO with road rash. Someone who's been there and failed. Someone who's been through all this. Now I think I qualify. Now I've got road rash. Previously I was a junior, I was green.
What are you doing now?
So if I could, if I had the money, I would go full-time in this new company that I'm building. But I need the money. So I'm going three days a week now at this new company, which is good, and it will keep me alive.
And then, as I build this new thing if it goes well, I'll just slowly reduce my time. And if I go and raise money, if it goes well enough and I want to raise like a decent lump of cash, then I'll go full time on this new thing.
What advice would you give to someone who is just starting out?
I think the biggest danger right now, is that you use ChatGPT to make a bunch of code you don't understand.
So I'd say if you want to be a programmer, make sure you're excited by it. Like it really genuinely interests you, and make sure that you understand it. Make sure that you learn a language or build an app where you have to actually really understand what's going on inside the computer.
It doesn't have to be C. It could be anything. Then build a nice app, or a website or something, and you'll be more likely to get a job. And then you are also more likely to do well in that job because you're going to understand what's going on rather than just relying on Stack Overflow or asking ChatGPT to do the job for you. That's my advice.
You don’t like ChatGPT?
Uh, it's useful, but I think people mistake it. People think it's this amazing, like conscious AI sort of thing or almost conscious AI.
In reality, it's the best auto-complete that's ever been created. You feed it a bunch of characters and it predicts the most likely best character, with some variance.
And because it's been trained on so much data, it's really impressive, but it also makes stuff up. There's a whole load of really bad code out there, which has gone into its training algorithm. So it will give you loads of nesting, it will give you really poor performance code, it will sometimes give you code that it confidently says will work when it doesn't. Random libraries that don't exist.
So it's really useful. It's a really powerful tool, but I think people use it or depend on it too much.
Do you think these AI systems will replace developers like us?
I think eventually, yes, but I think everyone thinks this is going to happen next year. I think in reality, this could take decades. It's really cool, it's really interesting, but the technology that's made ChatGPT has been around since 2017, so six years. And the technology that's based on, has been around for 30 or 40 years. The only difference between what we had then and what we have now is the scale.
I think to get to the next level of actual AI, like a more intelligent, maybe even conscious AI, I think we need to have a fundamental breakthrough. I think we need to be using fundamentally different technologies. It could be that we need a breakthrough in quantum computing, or even in biological computing.
Famous last words, I could be really wrong, but my gut feeling is that this technology we have is going to have a natural limit and we're not far away from it.
That’s a wrap!
If you have any questions for Rob, reply in the comments below.
See you next time ❤️