My Skills

Python

Been using Python for many years, for quick scripting, data science&analysis as well as API and heavy backend applications.I am very familiar with advanced features and know when to not use them. I have used most of the popular libraries at one point.I keep to a functional programming style but can build complex classes when necessary. Know how to speed up Python code and analyse the performance. Most of all I am very adept at using the pytest framework to build well tested and maintainable code.

For dependency management I have expertise in the use of use PyEnv & Poetry unless the project is just simple scripting where I would use venv. For any sort of APIs I tend to use FastAPI with AsyncIO where I am quite proficient to build a well structured easy to maintain API that performs superbly. Any work with databases would be done via repository patterns and usingwell known Python libraries such as SQLAlchemy, py-redis, pymongo, etc... Whenever data analysis of any sort is needed I am well versed in NumPy as well as Pandas (which I am not a fan of) combined with Plotly or Bokeh for visualisations. When it comes to Machine LearningI use PyTorch as I find it easier and more Pythonic. I am a huge fan of Pytest where I love to build a good fixture structure and stomp flaky tests.

GitLab

I am one of the biggest fans of GitLab. I love using it, I find the UI super intuitive. And I enjoy working with the CI/CD that offers amazing opportunities and power. I am also very adept at using its webhooks and API to integrate it with any tool and build awesome automations.

My love for GitLab comes mainly from how well structured and easy to use the CI/CD is. I love constructing and optimizing very fast tests and build pipelines. I like creating a safe and standardized release processes. Working with multiple types of runners and SSH makes it super easy to execute different jobs in different environments. Additionally, building scheduled and manual pipelines for various jobs is a breeze. As well as using GitLabs own package repository for Python and other internal library packages as well as the Image Registry they provide.

Docker

Docker is definitely a must tool for anyone who does any type of coding or collaborating. I have an in-depth understanding of Docker which allows me to help optimize and debug any issues. I especially enjoy when its time to optimize the build speeds, orreduce image sizes. For most projects combining this with docker-compose is enough to build any mid-scale application long before k8s is needed.

DataDog

I have always loved Observability and the value it brings to developers & non-tech alike. But with DataDog its fun as well.I love the ease at which logs can be parsed, searched and structured in the platform. They truly make identifying and analysing a situation a breeze.On top of that instrumenting an application to have metrics and events makes monitoring mission-critical applications easily achievable.

One of the projects I am most proud of with DataDog is a series of dashboards I built. These dashboards monitored & visualised the business logic of some backend applications our team was responsible for. This allowed the non-tech business teams to easily understand what was happening, react without needing to interrupt developers work or have the significant delays from before.The second project I enjoyed building with DataDog was the Quick Health series of log analysis that allowed us to quickly know what was deployed whereand do a series of post-deployment checks which allowed our team to deploy new features on a daily basis.

Atlassian

Jira & Confluence were the first collab tools I encountered when switching to the software industry.Since I like well structured things with a neat UI I immediately loved them. Confluence is still my favorite way to build documentation. It makes it incredibly easy to structure and collect information, very easy to collaborate and embed many nice visual elements. JIRA on the other hand I love mainly for its super intuitive Automation which makes for a great way to build complex workflows in no-time

The project I am most proud of was at my first company where I built up the much needed documentation for our team and even to this day I still get told its used and my name remains mentioned. The second project I am proud of is a release process which made it easy for developers to release our several microservices while the automation ensured. GitLab and Datadog were correctly updated and tagged for easy debugging and on-call support.JIRA was correctly updated and the correct Slack channels were informed for organisational transparency what was built where and when.

Kubernetes

I have a good understanding of the fundamental structure of k8s, from the control plane, worker nodes etc... As well as the key building blocks such as pods, services, jobs, replicasets, volumes and so on. I am also of the firm belief that only organizations at scale need k8s and others could use simpler setups.

I have had the privilege to work with k8s as a power user while a competent DevOps team managed its internals. So I mostly interacted with it trough Helm templates and charts, as well as ArgoCD which made my life much easier.

Kafka

A distributed streaming platform capable of handling trillions of events a day.

Linux

While I am not a Kernel expert, I have been a Linux power user for most of my adult life. I know where things usually are, whats worth modifying and how to debug certain issues. Bash scripting might not be fun, but sometimes its the easiest thing to get the job done.

AWS

I have used AWS as our cloud provider and am mostly familiar with the Kubernetes engine, image registry and DynamoDB.

Hetzner

Hetzner is currently my go-to server provider because its cheap, runs well and is extremely easy to use either via GUI or Terraform

Redis

Redis always comes in handy, whether it be some caching for a distributed service, or an easy way to share data between services. A project I am most proud of was using redis as the middleware from a websocket stream to a on demand REST API.

MongoDB

MongoDB is my preffered databse for documents and unstructured data. While I think SQL is a better lookup language than the mongos JSON based I very much like the performance and ease of use in Mongo.

Postgres

I have used one form of an SQL database in almost all of my work. I am familiar with good relational database design, normalization. As well as hard learned lessons about migrations and the importance of backups.

Angular

I learned a bit of frontend when working on a couple projects, and Angular is my favorite so far. I like when things are well structured and modularity can be easily achieved.