Presented by DataRoot Labs

Developing an industry-transforming AI solution is an incredible competitive advantage — but pursuing that goal requires real business transformation, along with a significant investment in talent and resources. Major challenges remain: the skills gap between demand for these very specialized roles and the actual pool of available researchers is still vast. And real AI innovation in business comes from a very particular combination of fundamental and applied research, the difference between pure science and goal-oriented, solution-focused work.

An AI R&D (research and development) lab is purpose-built for this challenge, merging the best of academic research with industry-centered goals, in a team with truly multidisciplinary AI skills, concentrated on a business partner’s specific AI objectives, says Max Frolov, CEO and co-founder of DataRoot Labs.

“AI R&D merges the real-world requirements of the technology sector, with a particular emphasis on pioneering new technology methods, and practical application technology concerns, and offers experimental ways to solve unsolved challenges,” Frolov explains. “AI R&D data scientists are dedicated to working on the big challenges — the ones that are changing the landscape, pushing boundaries, and ultimately developing real IP that adds a competitive advantage and results in a higher valuation of the business.”

Almost every company in Silicon Valley has their own AI R&D lab, whether internal or offshore — Frolov explains that companies like Amazon, Grammarly, Ring, Snapchat and DataRobot have already established an R&D presence in Ukraine, while the majority of the enterprise market intensely focused on internal AI R&D experimentation.

However, an external AI R&D can extend the frontiers of a company’s technology, no matter what size it is, and futureproof the business in a fraction of the time, resources and dollar cost of building an internal team.

Here’s a look at how those external AI R&D centers function, how they can help companies of any size develop innovative products and technologies that can transform markets, and how to establish a successful, lucrative partnership.

How R&D centers work

An AI R&D lab is an interdisciplinary team of machine learning (M&L) researchers and engineers, MLOps and natural language processing (NLP) professionals. A partnership offers a collaborative, agile and affordable way to co-build next-generation products and services with expected, but sometimes unknown or surprising, outcomes.

It’s able to apply relevant state-of-the-art AI solutions to new questions. Instead of being limited to working solely on product plans, researchers have the freedom to conduct experiments and delve into uncharted territory. They’re particularly effective when focusing on a niche area, whether that’s a specific AI technology like computer vision or generative AI, or industry- and market-specific questions.

An R&D lab can operate as an extension of a company’s core team, or as its primary AI partner, offering companies a collaborative development process for long-term projects that focus on problems demanding very deep AI expertise. A lab will often tap the expertise of local universities and other players in the ecosystem in order to ensure they’re up to date on the latest research, best practices and strategies.

“It’s very different from the consulting model, where you tap into high-level expertise and are charged a lot of money per hour for big-picture thoughts,” Frolov says. “With an R&D center, you have a team that works continuously on your projects to crack open unresolved tech challenges and deliver results.”

The work typically starts with analyzing the technical task and laying out the market and research landscape around the client. Architecture planning, and then sourcing and analyzing data from the client and elsewhere comes next, in order to train the model and test their hypothesis.

The hypothesis aims to answer two questions: Can the technologies we have today solve this particular problem, and if so, how can it be done? From there, the team develops a minimum viable product, and work continuously to improve the initial model.

“The important thing about the AI R&D setup is that although each client has their own dedicated team, it doesn’t prevent the overall lab from talking to each other and tapping into disparate teams of experts inside the lab,” Frolov adds. “If anyone is working on a task that requires, say, MLOps expertise, those teams talk to each other, do cross-checks and exchange knowledge. That results in more innovation to push the boundaries of what’s possible today.”

The advantages of an external AI R&D facility

A key advantage to an external AI R&D facility is solving the resource challenges that the industry faces right now. With the financial investment required and competition for a very limited pool of experts, a lack of in-house resources continues to be a critical obstacle to launching an AI strategy.

“And no matter how much you’re willing to pay, there’s still a shortage,” Frolov says. “If you set up your AI R&D center somewhere offshore like Ukraine, not only can you tap into a vast and varied engineering pool outside your usual boundaries, but that engineering pool is typically far more balanced in terms of seniority level and comfortable for R&D team pricing     .”

It’s also a far more stable source of talent, he adds. In-house engineers frequently leave jobs in which they feel stifled or unfulfilled, or because there’s a better opportunity elsewhere. Poaching, of course, is not uncommon. But in a lab, they’re in a tech-focused, entrepreneurial research-based environment that offers true collaboration. And they’re working on big, industry-shaking challenges, where they can actually develop their skills, expand their expertise and wealth of knowledge and advance their career while working on thought-provoking projects.

For the clients, that means a long-term R&D partner, with engineers that you know well and trust, who are dedicated to your cause. You can create a long-term relationship committed to advancing and continuously iterating on your AI goals.

Tapping offshore talent

One of the best reasons to go outside Silicon Valley specifically is the cost. For instance, the salaries for engineers in central and eastern Europe, India and South America are significantly lower compared to North America, so the cost of development is significantly lower. These countries are also investing heavily in their talent pool, not only partnering with universities, private and public research organizations and the like, but supporting the next generation of skilled data scientists.

For instance, in a location like Ukraine you’re tapping into a large technical talent pool, but it’s not only about the numbers, it’s about the quality of engineers. The country has a long legacy of scientific education, Frolov says.

“We collaborate with the best local universities that offer tech education, including Kyiv Polytechnic Institute, the MIT of Ukraine,” he says. “With Kyiv Polytechnic we’re establishing a master’s program in AI. And our own free online school, DataRoot University, currently has about 6,000 students registered.”

DataRoot assists students with the practical application of their technical knowledge as they advance. In teams of three to five people, students work on AI startup project ideas for six months, with an assist from DataRoot, which will support successful projects into completion. 

“For us it’s a mission — to pay it forward and grow the field,” Frolov says. “And our goal as a company is not only to push the boundaries of technology, but also to put Ukraine on the map of the AI ecosystem out there. But one of the top reasons to hire in Ukraine now is that by working with Ukraine you help a country that’s been invaded by a hostile nation — but is still working on creating tomorrow.”

Learn more here how AI R&D-as-a-service can boost innovation to seize competitive advantage.

Sponsored articles are content produced by a company that is either paying for the post or has a business relationship with VentureBeat, and they’re always clearly marked. For more information, contact