How One AI Developer's Agency Helps Companies Navigate Change

Industry

Where data and ideas come together

While Ortega started working in machine learning as an independent professional on Upwork, he eventually decided he was more interested in building a circle of talent around him. “Great things need a great team,” he said. “I saw projects I couldn’t do by myself and I started to feel that I wanted to share my work with a team, to be able to reach more complexity.”

He launched Eidos.ai, a development studio that specializes in machine learning solutions. The name Eidos was inspired by a concept from Plato’s philosophy, which describes a world of ideas that coexists alongside the physical world. “It’s a split between the reality of the physical world—sort of like pure data—and ideas, or the meaning that humans bring to that data,” Ortega explained.

Eidos.ai helps connect data and ideas for businesses that want to adopt AI and grow.

“A big portion of the customers we work with are startups trying to develop an idea that has become possible with new applications of AI,” Ortega said. “They generally make a proof of concept with us, then pitch it to investors so they can start growing the product they’re dreaming about.”

Another segment of Eidos.ai’s customers are existing product teams, companies that are already up and running but want to explore how AI can help them be more productive and scale.

Small agencies as innovation engines

Many companies see the potential of AI to help them work better. But they don’t always have the resources to figure out how to generate that impact. Leveraging an agency like Eidos.ai is a way to try new things without diverting an in-house team from other priorities.

“Doing something in-house doesn’t just take engineering time, it also puts pressure on project managers and other resources,” Ortega said. That makes it difficult for companies to create a good environment for innovation—if they even have access to the skilled talent they need.

“There are a lot of companies seeing this AI trend and starting to think, ‘Hey, we should think about this,’” Ortega said. But AI often requires a great deal of hands-on innovation, and he said a company may not know whether something will ultimately be worth pursuing:

“What’s a better option: Hiring someone who has to learn about your company as well as their new position, or working with an agency who’s done this before with other companies—maybe even ones similar to yours—who understands the challenges and current trends impacting AI projects?”

An agency can provide a development capacity boost from the beginning, then gather the information needed to decide whether to move forward. “As a partner to an existing in-house team, we can spend time exploring an idea and offering potential solutions. It's a matter of complementation.”

Problem-solving for a new sports application

One of the challenges and rewards of working with AI is addressing problems that haven’t been solved before. “A lot of the time, with software development, you can find and build on existing solutions. With AI, that isn’t always the case,” Ortega said.

One of Eidos.ai’s client projects has required a pipeline of different AI models. “We are developing a system that extracts information from sports videos,” Ortega said.

The system detects the position of players and collects a wide variety of data points that are then converted into performance statistics. “This information can help teams decide, for example, that a player should change position because he’s playing better on one side than the other,” he said.

“There have been a lot of scenarios where the problem we were trying to solve had not been talked about on the Internet,” Ortega explained. “We had to create our own AI models to address novel problems—maybe based on something already established, but adapting it in ways that weren’t used before. It’s been very rewarding to see our own solutions at work.”

How companies can think about adopting AI

As generative AI continues to iterate and expand, businesses of all sizes are wondering how to use the tools and technology that have emerged with it. Ortega said there are a couple of concepts that can help frame expectations:

“If a human can do it, maybe AI can do it. If a human can't do it, AI probably won't be able to do it either. You can train an AI to do something similar to how you do it. But it’s going to be hard to get it to do something better than you can.”

A good way for companies to start adopting AI is to consider activities that are done repeatedly. Ortega said these are often everyday processes that take a great deal of time, are fairly manual, and always need to be done the same way.

Before you can launch any AI implementation project, however, you’ll need to define a data strategy.

Which comes first: Data or AI?

“Data is the key because AI lies on top of data,” Ortega said. A data strategy should address questions such as:

How are you going to collect or access data?
Who will own and govern the data that’s collected?
How are you going to get data that applies to different situations?
How are you going to ensure or validate data quality?
Where will you store the data?
How are you going to ensure data security?

For some projects, the data requirement creates a deadlock. “To make good AI, you need data. But to get good data, you need to offer customers something good,” Ortega said. “It can be hard for companies to get inside that loop.”

That means the goal for many AI projects is to start something as soon as possible. “Get the first customers, get the first information feeds,” Ortega said. “After that, it’s a process of improving the quality of what you offer. The data will help you do that.”

Staying up-to-date as AI continues to evolve

The AI industry is evolving at a rapid pace, which makes keeping up with new developments, trends, and best practices no easy feat. Ortega offers a few suggestions.

1. Make education a team effort

Every two weeks, someone on the Eidos.ai team picks a topic to research, then presents what they’ve learned to the team. They also use Slack to share information and links that might be of interest.

2. Make time for social media

Being active on networks like X, formerly known as Twitter, is an important part of Ortega’s professional development. “Just seeing posts on your feed is not going to give you a very in-depth idea, you have to open the reports and study the information,” he said. “But even scrolling helps you stay in touch. You might see a term one day and think, ‘Hey, that’s new.’ Then a week later, when you’ve seen it 20 times, you might decide to check out what’s going on.”

3. Immerse yourself in AI development

While running an agency means Ortega doesn’t create code every day, he said building your AI skills and working with it regularly is the best way to stay in the know. “The things we develop for clients and new releases have to be very fresh,” Ortega said. “Being there on the frontline developing stuff is the best way, I would say, to stay updated.”

Important skills to sharpen, especially for a career in AI

One skill Ortega thinks is important for any professional—working in AI or not—is learning how to use AI in your daily work. “Companies need to give people the opportunity to use it. But then it’s on everyone to identify situations where AI helps them work better,” he said.

Ortega said there are a couple of skills he looks for in a developer. The first is the ability to think abstractly. “It’s hard to find but being able to look at the high-level picture, without getting stuck in the details, that’s very important,” he said.

Another essential skill is communication. “Being a good developer isn’t just about developing—you need to be able to relate to other people on the team as well as the customers you work with,” Ortega explained. “It’s important that you can understand what other people are saying and that you’re able to give your own point of view.”

Surfing the AI wave will take self-reflection

Ortega admitted that owning a development studio that’s focused on AI doesn’t mean he’s completely at peace with it. He sees a wave of change coming to the software development field and believes surfing that wave successfully will take individual introspection.

“I think we need to ask ourselves: What do I offer to society in general? What am I good at doing?” Ortega said. “Question both the things you do because you have to and the things you do because they have value. AI is going to eliminate many of the tasks we do because we must and, in an ideal world, give us more time to do the things we want to do. It’s going to be a challenging time. But it’s also going to be cool to be part of it.”

‍