Have you ever asked yourself what is the limit of artificial intelligence (AI)? Can a machine learn to program all the innovative solutions that we use? Imagine if there existed software capable of proposing to programmers the exact code they need. We have great news: it is no longer necessary to imagine it, it exists, and it is called Augmented Coding, and you will be able to find out about its potential on September 17 at Converge+.
As the event date approaches, we invite you to get to know how this innovative solution, that promises to transform the future of programming, arose and became a reality. Discover the history of Augmented Coding told first-hand by Haldo Sponton, VP of Technology and Head of AI Development at Globant.
How was Augmented Coding born?
Augmented Coding arose from experimentation. At the beginning of 2019, we were “playing” with linking natural language and programming language, for example English and Python. After a month and a half, we arrived at the first result, a small “engine” which could ask questions in natural language and it would return the relevant code as a response to that question. We had managed to link data between two worlds!
On February 5, 2019, in New York, I showed the “engine” to Martín Migoya, Globant’s CEO and Co-Founder. I am never going to forget that day. We spoke about its potential and he told me “this tool could perfectly understand programming languages, help programmers in a lot of ways, and completely transform the way in which we make code.” That was a very powerful sign that we were on the right path.
How did Augmented Coding evolve and develop?
Starting with Martin’s vision and validation, I had all the support to set up a great team for research and applied development. Everything was very new and needed a lot of work before we could transform it into a product.
Within three or four months, we had our first MVP or solid proof of concept, which supported more than one programming language. During the second half of 2019, we dedicated ourselves to developing it, improving its interfaces and trainings, and developing its functionalities.
At that time, we realized something fundamental: the MVP not only responded to questions with code fragments, but it was also capable of performing the inverse operation. It could receive a code fragment and explain the operation to us. In other words, it automatically generated documentation for the code. So those clients with enormous undocumented code bases now have the possibility of having them transformed with original documentation immediately added to them. With that the code becomes understandable and maintainable. It’s amazing!
What challenges have you had and how did you overcome them?
One of the greatest challenges of the project was getting training data. AI algorithms need to learn, and for that, you have to give them large amounts of data and hours of training. We had to go out and search for open, quality data that could be used for training. It was not as easy as it sounds!
In addition, we constantly had to train the team to work on these models and artificial intelligence approaches, because we also needed to specialize in state-of-the-art natural language processing technology. Today, the team has 30 people, is constantly doing research, improving the product’s functionalities, and looking for new functionalities.
What impact does Augmented Coding have on how programming is done? And what opportunities does it provide for companies?
We are revolutionizing the way in which we develop software. This will help developers program in a faster and more agile way and code with fewer errors. This will help our clients be significantly more productive.
Today, AI is accelerating software development. With Augmented Coding, we are compiling this intelligence to place it at the service of developers working for clients.
To our clients, code is gold. For this reason, the product has to be installed and the code processed in a secure manner. No complications can occur, and it has to be able to provide developers with fluid and intuitive access to the tool.
In addition, each functionality of Augmented Coding (self-completion, automatic documentation, semantic search), depending on the client’s use, can be trained, and may evolve for what each client needs. The potential that Augmented Coding has is tremendous.
Why did it become more relevant during the pandemic?
When confronted with a distributed work scenario, it was essential to support our developers, who no longer had the in-person support of a leader or technical support staff.
We achieved this with Augmented Coding, which worked as an assistant who automatically responded with the code that individuals needed. The tool quickly became the first point of consultation.
In the future, remote work will be more and more common. For this reason, it’s important to have technology that helps teams collaborate.
Augmented Coding is based on collaboration, it learns from the operations that everyone performs with the tool. Whenever valid relationships between questions and codes are established, they are saved as useful links. When these relationships multiply, a knowledge map is created with an extremely large amount of available information. This ends up impacting speed, efficiency, and productivity.
What expectations do you have for the Converge+ event?
First of all, the most important objective is to introduce the tool and communicate its immense potential. With Augmented Coding, we are really going to disrupt the software industry. We want to share this message with everyone attending, and show them this new vision of programming.
We also are extremely excited by the incredible speakers that are coming to the event, and with whom we hope to talk about these ideas.
| CONVERGE+ On September 17, Globant invites you to participate in Converge+ Reinvent the Future: Organizations Strengthened by AI At the event, Augmented Coding will be introduced, an AI-driven solution that will forever change the software development industry. Some of the top speakers who will participate in the event are: • Steve Wozniak (Top Technology Speaker, Innovation Expert, Cult Icon, Co-Founder of Apple Computer, and Bestselling Author) • Martín Migoya (CEO & Co-founder, Globant) • Haldo Sponton (Head of AI Development) • Michael Feathers (Founder and Director of R7K and author of “Working Effectively with Legacy Code”) • James Taylor (Creativity Speaker, Award-winning Entrepreneur, Founder of C.SCHOOL) Come get to know this new form of programming and discover how it can change the present and future of businesses! |
| HALDO SPONTON VP of Technology Head of AI Development Haldo is an eager technology and math lover, with vast experience as a teacher and researcher in signal processing and machine learning in universities in Uruguay and France. His primary focus is the strategy of digital and cultural transformation of companies based on a data driven mindset. He is a VP of Technology and Head of AI Development at Globant. He’s the leader of the Augmented Globant team with the objective of implementing and realizing the idea of Augmented Organizations in order to transform Globant and organizations. This team develops tools based on Artificial Intelligence that transform the way we code, the way we find the best talent, the way we design, the way we organize anda access all our information, and the way we keep our culture alive. Haldo has an Electrical Engineering degree, a Master in Sciences degree in Applied Mathematics and a Tech MBA degree. He has been keynote speaker in different tech events, and he’s the co-author of the book “Embracing the power of AI: a gentle CxO guide to Artificial Intelligence”, in which the impact of AI is analyzed both from a technical and business perspective, among other scientific publications. |
