An incredible year - 2024 in review
- Victor Hugo Germano
- Dec 18, 2024
- 4 min read
2024 a year of learning and reflection!
This year has been incredible, to say the least. Since my announcement that I would be leaving Lambda3 , I have taken a sabbatical from the corporate world as a way to rest, but mainly to deepen my knowledge and experience in other areas, more connected to what I hope to do in the future, as well as a path for my next stage in life.
I recognize how privileged it is to be able to stop working after a lifetime of work and pursue other activities, and how this is exclusive to a very small group of people. I hope that one day we can have a society in which no one has to work to the point where they can't keep up with life .
I started a year of travel and I was here recording the whole process.
In the meantime, I ended up really delving into Artificial Intelligence, from its conception, but mainly in the more political, ethical and speculative contexts of AI .
Content and Communities
There were some lectures
I'm no Longer a CEO. What's the Next Challenge of the Decade? on Afroya Techhub
Class at IFSC - From zero to one hundred in 10 years
There were some podcasts
Devs 40+ telling a little about My story
Businessmap Episode: Fundamentals of Modern Management
Pogcast Episode: About Agility and Development
LTP Episode: The implications of AI in the dev market
There's Time to Ask an AI Question Episode
Agile Popcorn Episode: Does Agility Have to Do with Artificial Intelligence?
In addition, there were many posts here accompanying my study on AI and the Market, which I highlight:
It's a return to writing that I've been trying out for a long time, and which is a reflection of the time I currently have available 🏖.
This was also the year of Bluesky in the world, and I have been updating this social network much more than the others. You can find me at victorhg.com . I have also been sharing a lot of my experience with Obsidian as a note-taking tool , and I am converted!
Master in Applied Artificial Intelligence
I completed the first semester of the Master's in Applied Artificial Intelligence at the University of San Diego. I only have positive things to say about the work and the content. Although it is an initial content that could easily be replaced by a course from Udemy, Briliant or Alura , there is a greater emphasis on the connection between the participants, and the indication that everything we are learning needs to be directed towards the industry: at the end of each discipline, a presentation and mini project that solves a real problem is required.
There were basically two works carried out this semester, which are also published on my GitHub: A statistics exercise with Magic the Gathering cards and a model that recognizes PPE in Civil Construction images :
Price Prediction for Magic the Gathering Cards
We decided to use multiple regression and classification techniques to enable us to predict prices based on the categorical data of the Cards, using Linear Regression and Ordinary Least Squares (OLS) Regressions. We will then employ a Self-Organizing Map (SOM) to aid in clustering and visualization. Finally, using Random Forest Regression, we will compare some of the results obtained from the SOM with our own manual statistical evaluations and lay the groundwork for building confidence in machine learning models for multidimensional datasets. Many fields can benefit from the application of simple machine learning algorithms such as SOMs.
Given the maturity of the Magic: The Gathering card market, we have identified an open-source dataset that includes current (or relatively recent) market prices. This comprehensive dataset will be our ground truth, rather than aggregating data from multiple markets.
Secondary markets are influenced by many factors when pricing a card, many of which are external to the data used in this article and could be used in future research. Card synergy during play, deck strategy in tournaments, exclusive art, and celebrity endorsements can all affect pricing and will not be used in this analysis.
A real application in an industry that is expected to grow to over $10 billion in the coming years.
The technical document describing our activities , as well as the presentation, are publicly available to anyone interested.
Recognizing people using PPE in Civil Construction
Construction workers are at a disproportionately higher risk of fatal and non-fatal injuries compared to workers in other industries. The absence and improper use of personal protective equipment (PPE) are leading causes of fatal and non-fatal injuries on construction jobsites.
However, manually monitoring PPE compliance is quite complex, involving daily supervision of a large workforce. An alternative is to utilize the latest computer vision (CV)-based solutions and frameworks to determine whether workers are using PPE in accordance with safety protocols.
Video monitoring also sparks numerous ethical conversations about the effectiveness and legitimacy of recording professionals in the workplace at all times, and this was certainly a discussion we had. Overall, the tooling for creating, training, and testing CNNs is quite mature, and can have a real impact on supporting occupational safety professionals.
A very cool project to carry out, and it is also available on github .
For those interested, the technical document and presentation are also available.
Around the world in photography
This year I visited 32 countries, including a wonderful trip to Asia. It was transformative to spend so many weeks in China, South Korea, Japan and Thailand. It was my first time in all of these countries, but it will hardly be my last!
On Instagram I shared countless photos of these places, and I continue to share my learning in photography.
I wrote some reports for those who want to know a little about this journey.
Sabbatical Project - Q2 Review on France and China
Sabbatical Project - Q3 Review on Korea, Japan and Thailand
An incredible year of discoveries and learning! Bring on 2025!
Comments