Simple DevOps for Data Scientists, Business Process Automation with NoCode and Low Code Tools

Hi, I am Chris and I build Business Process solutions with NoCode and actual code.

My main motivation is to free up brain capacity and time to work to be able to focus on problems that matter.

My method is to deploy extensible systems to automate data collection, transformation and analysis to facilitate decision making for executives and business owners.

I use well maintained Open Source Software to avoid vendor lockin, to minimize custom code, and allow for scale. All tooling that I built is portable can run locally on your laptop/desktop, in your local network, as well as in the cloud on single servers. If needed the same code can be deployed to scalable cloud architecture.

I'm also a big proponent of being able to self-host solutions based on Open Source Software. While Cloud Providers like Amazon Web Services, Google Cloud Platform, Heroku or Microsoft Azure are providing an insane amount of services, they also come with a strong lock in effect. For the more privacy conscious there are alternatives now thanks to a lot of battle tested open source solutions:

My core technology stack consists of Python, Docker, Ubuntu LTS, NGINX and Relational Databases (SQLite, PostgreSQL).

Current Focus

DevOps for Data Scientists: how to build and share your AI/ML Models with the world

• Deploy Meltano as "ELT for the DataOps era". It allows to have a single source of truth for all your data sources. From there we can build dashboards (Self-hosted or as a service: redash.io, Metabase, Microsoft PowerBI, Tableau), integrate AI/ML models (Tensorflow, spaCy, scikit-learn , PyTorch, Hugging Face, Keras, Cloud Provider Solutions like AWS Sagemaker, Google Cloud AI, IBM Watson) or trigger alerts to take action based make informed decisions.
• Applicable for small to large scale operations who find value in unifying their data and be ready to apply Machine Learning and Artificial Intelligence models.

API Development

I built my own APIs with FastAPI. They are fast, come with interactive documentation out of the box, and adhere to OpenAPI standards for fast implementation.

Workflow automation with n8n.io

Deployment of n8n.io on your local machine (PC, Mac, Linux), local network (Raspberry PI), or VPS in the cloud.

• You get your self hosted automation platform, with a graphical user interface, which can be extended with custom nodes.
• Create your own custom workflows.
• Periodic Backup of your workflows.

Blending "real code" with NoCode tools and hosted Application Development Platforms

The main and common benefit of these platforms is that they handle authentication and authorization for us. Depending on the use case and ressources, we can leverage a NoCode platform to rapidly built out prototypes or go for a more custom approach. We extend the limitations of these platforms but integrating our own APIs to trigger, pull, update information.

• NoCode: Bubble.io for rapid prototyping of real web applications with visual programming. Great for End User facing platforms, as well as internal platforms.
• Application Development Platform: For rapid "real" software engineering. Great for all kinds of web applications.

Why is Website still so empty?

The layout is the work of a weekend project (see Twitter thread). It's part of the #buildinginpublic and #learninginpublic movement. It's built on the Django Web Framework which I haven't had the opportunity yet to learn.