An A.I. vision for the future of learning to code
An A.I. vision for the future of learning to code
Cubode wants to use no-code to teach non-coders to code
PreSeed Now brings you an in-depth profile of a different B2B or deep tech startup every Tuesday and Thursday. Subscribe to get it straight to your inbox.
Hello there,
How does coding education change in a world where A.I. can help us code?
Today’s startup is exploring that question with a product aimed at data analysts, but with the potential to expand into other verticals.
Scroll down to read all about Cubode.
Meanwhile, following a positive reaction to Thursday’s edition about the state of pre-seed investing in 2023, we’re collecting ideas for other feature pieces in the future. If there’s something specific you’d like to see covered, drop me a line!
– Martin
💡 Investors: let’s talk due diligence
When you invest in a startup, you need to be sure their tech–the true value behind the deal–is everything they claim it is.
PreSeed Now is sponsored by thestartupfactory.tech’s Technical Due Diligence offering
This service goes much deeper than the technology itself, taking in leadership, technology strategy, product and roadmap, people, engineering practices, and commercial analysis
Cubode has an A.I.-focused vision for the future of coding education
Much has been made of the latest generation of A.I.’s ability to write code. But in addition to transforming the role of existing software developers, there’s potential for this shift to help empower those who are learning to code.
Case in point: meet Cubode, a product helps data analysts create advanced analytics quickly, while teaching them how to code, with the help of A.I.
Co-founder and CEO Rebeca Garcia explains that the target market is any analyst who uses data on a daily basis, in any industry, but who lacks the coding skills to create advanced analytics like interactive visualisations or machine learning models.
Cubode offers a graphical interface that allows users to create these advanced analytics, generating Python code in the background.
“They get a full explanation, step-by-step, of whatever they've created graphically,” says Garcia. “They can create graphically, get the outputs within the canvas, get the Python generated code, and then reuse that however they want. They can run it locally if they want to.”
And Cubode can translate the finished code into whatever coding language its users might need for other purposes.
“They also get a full explanation so they can learn about what they've done,” adds Garcia. “That unlocks upskilling to coding at scale, because analysts can learn about what they need, specifically for their business use and with their own data.”
These explanations are generated by large language models (LLMs) but Cubode wants its relationship with these much-hyped A.I. models to go both ways. The startup is working on an API that will allow LLMs to draw on Cubode as a source of information, helping them generate accurate code examples, something they don’t always do from large, unfocused training sets.
When automation isn’t enough
Hearing all of this from Garcia, I had a perhaps obvious question: if Cubode can do the job for them, why should its users learn to code?
Garcia says the idea of Cubode isn’t to be able to do everything for its users. Instead, it lets them realise their ideas via a range of ready-to-use templates.
Users can see how these templates translate into code that is based on their own data. From there, they can begin their journey to coding their own custom data applications.
The templates are made up of ‘cubes’ that represent Python functions, such as being able to read a CSV file, or grouping a data set by cities. The templates make it easy for the user to see how these different cubes are connected to each other, and to understand what the code is doing with their data to get the result they want.
And Garcia is a firm believer that A.I. code generation cannot replace human coding.
“A.I. predicts and learns from past code, but humans are the innovators. New ideas happen when you don't waste time doing the same things over and over again.
“The true value that Cubode brings to the table is that it allows innovation for people who didn't have access to innovation before.”
Progress so far
Liverpool-based Garcia co-founded Cubode with her brother, Munich-based David Garcia Corrochano. After beginning work on the product a couple of years ago, they first announced it in November last year. She says since then they have built a waitlist of more than 600 potential users across 90 countries.
Garcia says they have learned a lot about usability from the first beta tests they conducted. These led to them adding an onboarding process that explains the platform, and a better interface for template discovery.
She says they are currently working on improving usage metrics, but she’s encouraged to see that users are spending up to 44 minutes with the platform in a single session right now.
A further two beta tests are planned ahead of Cubode’s official launch, which is currently scheduled for November. Garcia says they will attract users upon launch through an influencer marketing campaign on social media platforms.
Background
Garcia grew up in northern Spain, where she began her career in marketing, before moving to the UK to work for THG’s Myprotein brand as a commercial marketing manager.
She then pivoted her career to become a financial analyst. This meant she had to learn SQL and Python to analyse investment portfolios.
“That's when I understood the struggles of learning how to code. I had the commercial knowledge of what I wanted to build and what I wanted to extract from the data. But I had no idea how to extract the data and how to create anything analytically meaningful without the coding skill, so I had to learn how to code.”
Garcia says she brings first-hand knowledge of the problem they want to solve, along with the marketing nous to grow the business. Her brother and co-founder David, meanwhile, brings the tech expertise to build the product itself.
She says David first learned to code with Visual Basic on a 286 PC around the age of 10. He’s now an aerospace engineer by profession, with a specialism in A.I. and machine learning, and he brings these skills to developing the Cubode product.
Go deeper on Cubode
Read more on their investment plans, vision, competition, and challenges:
Investment:
Keep reading with a 7-day free trial
Subscribe to