EPAM is ramping up its training for generative AI. The company has launched strategic partnerships with Anthropic, OpenAI, and Google, through which it trains and certifies its specialists to work with Claude, OpenAI, and Gemini models.
EPAM is ramping up its training for generative AI. The company has launched strategic partnerships with Anthropic, OpenAI, and Google, through which it trains and certifies its specialists to work with Claude, OpenAI, and Gemini models.
According to Stepan Mytish, Vice President and Head of EPAM Ukraine, within the framework of the program with Anthropic alone, the company has already trained over 5,000 certified Claude Code architects. The overall goal of the program is over 10,000 such specialists. EPAM officially announced the formation of a practice with over 10,000 Claude-certified architects.
«We have agreed with Anthropic to globally train 10,000 certified Claude Code architects at EPAM — over 5,000 have already been trained, and a large proportion of them are Ukrainians,» Mytish said in an interview with LIGA.net.
The company launched a separate program with OpenAI. According to Mitish, EPAM plans to train another 10,000 certified specialists to work with the company’s models within a year.
In its own partnership announcement on July 28, EPAM said it plans to certify more than 5,000 consultants and obtain more than 10,000 related credentials within the first year of its partnership with OpenAI. The goal is to help clients implement OpenAI models in regulated and business-critical environments.
EPAM also has a similar program with Google to work with Gemini. In early 2026, the company entered into a strategic partnership with Cursor.
Thus, EPAM does not rely on one specific AI model, but on the ability of its teams to work with different tools.
«We believe that the winners will not be businesses that rely on a specific model — OpenAI, Anthropic, or Gemini — but those that have teams capable of implementing different AI models into real, complex business processes. There is a shortage of such specialists in the world,» said Mitish.
According to the head of EPAM Ukraine, the development of AI has already changed the set of skills needed by IT engineers.
«While coding skills were previously highly valued, now most of the not-so-complicated coding tasks can already be performed by AI,» he noted.
Therefore, skills that are more difficult to automate are becoming increasingly important: understanding the client’s business, the subject area, and the ability to transform a request into a ready-made production solution.
The traditional division between specializations is also changing.
«If there used to be a clear division into frontend and backend, separate programming languages, then thanks to AI these boundaries are blurred. A high-quality programmer, even without knowing a certain language, can quickly start writing in it using AI,» explained Mitish.
At the same time, he said, the company is not only training programmers to use AI. Employees from various functions, including accountants, financiers, and client partners, are also receiving training.
EPAM is also rebuilding its own ecosystem of internal products. The company is introducing AI agents and chatbots that can reduce the time it takes to complete certain tasks from days or weeks to much shorter deadlines. One example is analyzing a client’s RFP, or request for proposal.
«We now conduct and prepare a response to a client RFP analysis, which previously could take a week, using AI,» said Mitish.
Another direction is software development using AI agents. According to the head of EPAM Ukraine, they allow you to shorten the development cycle and perform more tasks in parallel.
In such a model, one person can actually coordinate the work of multiple AI agents that perform different parts of a task.
Mitish believes that for the IT services industry, this is not just the emergence of another tool. «The era of artificial intelligence is more than the emergence of the Internet or mobile technologies, it is a fundamental change in everything, including IT,» he said.
Cybersecurity has already become a separate area. According to Mitish, modern AI models can find vulnerabilities in software products in minutes or hours instead of months or years.
«For example, a European client ran such models through their systems, found a lot of vulnerabilities, and turned to us to quickly fix them,» he said.
