Inspirational 2 years side-project with architects leading to a Python library

The Spark: An Impossible Request

Disclaimer: NO AI bullshit here. We are in April 2024, I am working in Mexico City and a friend of mine contacts me. She is an architect and she works in Merida. Her and her colleague would like to have a solution to analyze architectural cadastral documents. She told me that she called friends from Mexico and USA and they all say it is impossible.

Interested by the challenge I said: "Hold my beer" and I made a POC in 75h of intensive work. However, it was not enough. They had other requirements and getting to this was much more complicated. Complicated but still not impossible.

Building the Solution

As a Data Scientist, I started to develop a full solution (mostly Computer Vision) to read and analyze automatically specific cadastral documents. I decided to continue working on the project. Hours and hours. There was no money involved. Only challenge. Thanks to that I was able to do everything with the time I wanted. Putting all the passion in my side-project.

With time the project grew and grew. But I was prepared for this. I made every part of the code separable, isolated. 6 repositories:

• aitt-core: core backend + Streamlit app -> Docker image deployment on ECR
• aitt-core-deploy: service deployment on the server via Ansible
• aitt-symbol-clf: ML + API -> Docker image deployment on ECR
• aitt-symbol-clf-deploy: service deployment on the server via Ansible
• aitt-infra: Terraform on AWS
• aitt-platform: Vue.js frontend (private repo)

I finally came to a solution for what they needed in more than 625 hours of work. My work contract ended in 2025 in Mexico so I came back to France (I am French, yes I love pain - joke for bilingual).

Lessons Learned

This project made me realize several things: I am capable of doing great things. We all are. The only thing that matters is discipline. This is not some sort of personal development bullshit. This is an advice for life.

Einstein said: "I have no special talents. I am only passionately curious." Time is a powerful thing. With time and dedication you can achieve great things.

It made me more confident to create my own business and go freelance in the mid-term. Before that, I thought it was only for genius or men with already a lot of money.

The Outcome

In the end, I separated the generic image & geometry codebase to make it a Python open-source library called Otary. Now I have for free two projects that are really important in my life:

• Otary (link to the docs): an open-source Python library where anybody can contribute and I hope will change the future of developers for image, geometry and computer vision projects. I am working on it with love and care on my side.
• Architect Solution: a private solution with already a solid backend that I could sell in the long term as a SaaS for architects in France or in other countries (still a lot of work to be done!).

I hope this will inspire some other programmers here. Happy coding!