I started my career in software development in the world of medical devices. From a piano teacher who wanted to change careers to have a positive impact in the world, I couldn’t have gotten luckier. Working in a startup, especially a medical device startup, I ended up doing a lot more than my title described. On top of learning about software development, I also got to learn about quality management systems, non-conformances, CAPAs, audits, data privacy, information security management, etc. In terms of software development though, I feel a bit stunted, and I’m ready for a new challenge.
As I’ve been applying to jobs, I’ve noticed a big gap between my experience on paper, and my experience on the tools. I realised that in the world of Python developers, there are: AI developers who have PhDs in machine learning, scientific programmers with at least a master in biochemical engineering (or some other “real” engineering), or back-end programmers with at least five years of experience. None of these describe me, so I need to get closer to one of those descriptions if I want a job!
I used to just think of myself as a software developer, but what I’ve been learning while looking at jobs is that the position I held was unlike anything else in the market. In hindsight, thinking there would be a perfect mapping is naïve. Anyhow, more tertiary education is out of the question at the moment, therefore, I need to aim towards back-end development.
In my first and only interview so far, I was tasked with designing an audit log system. It was more complicated than I had originally anticipated (partially from me overcomplicating the design) and the discussion with the interviewers highlighted some clear gaps in my knowledge and practice.
I’m a big fan of the idea of learning in public but I’ve always found it difficult to do. What is the point of me recycling knowledge that already exists in a better form? I’m not adding anything to the conversation. What I can contribute is my experience and my recommendations. In an age where AI has become a thinking and learning companion, I think it’s necessary to get back to real human recommendations from people we can identify with.
Below is a list of resources I found useful for learning. I hope you get some use out of them too.
Roadmap: This was a useful starting point to figure out what conceptual knowledge gaps I had. I’m pretty sure this is AI generated, but it’s a good enough start.
When I realised that I didn’t know as much about PostgreSQL as I originally thought, I went deep on SQL, database design, and DBMS internals.
This course is a great little primer on SQL with some fun exercises. I had never actually studied SQL and only gathered information here and there so it was nice to have it all in one place where I could also practice. Note: I haven’t finished the entire course yet.
SQLite has great documentation about its architecture and how it differs from PostgreSQL. There is also this book about the internals of PostgreSQL. I’m not super interested in understanding the deep inner-workings of these DBMS but having an understanding of pages, how indexes work, or how queries are computed is useful knowledge to help make sense of some of the prescriptions you will hear for designing and optimising data storage and querying.
In terms of query optimisation and indexing, there is this website. The default indexing strategy is a B-tree (balanced tree), the theory of which you can learn in Knuth’s TAOCP Vol.3, but for other indexing strategies, there is a short and informative video by HelloInterview here. Sharding and partitioning are also concepts to quickly look into, but otherwise the basics of query optimisation just seem to be common sense (e.g., only select what you require).
I also learnt about different communication protocols; turns out HTTP is not the be-all and end-all of application layer communication protocols! Two protocols I didn’t know about before were WebSockets (which enable two-way communication between server and client) and gRPC (which provides a standardised format for calling functions from other microservices).
Theoretical knowledge will unfortunately not be enough though. I need to show that I can apply the knowledge and use the technologies. This is the part that I’m currently struggling with. Systems require things like Kafka, Redis, or Kubernetes because they have a certain scale. Companies need CI/CD pipelines because they have dozens of developers working on a project and need to deploy quickly. None of these constraints are present for personal projects, so I’m struggling coming up with ideas of projects that are serious enough to warrant these technologies but no so big that I spend the next three months working on them. I’ve considered contributing to open source (something I’ve wanted to do for a while) but I’m finding it difficult to find projects where I will be able to get experience in these technologies.
If anyone has any corrections for my noob brain, suggestions for how to proceed, or is in the same boat and would like to bounce ideas, leave me a shout at my full name (url of this website) [at] gmail [dot] com. In the meantime I will keep learning and applying, and check back in when I have more recommendations.
Header Image
Boats at Berck-sur-Mer, Édouard Manet (link)