EPAM Systems (EPAM ) said on July 28, 2026 that it has joined the OpenAI Partner Network as an Advanced Partner, and that it will certify more than 5,000 of its consultants on OpenAI’s models within the partnership’s first year, carrying more than 10,000 credentials between them. The engineering and consulting firm is standing up a dedicated OpenAI practice to wire the models into the applications, data and workflows of Forbes Global 2000 clients.
The certification commitment is the countable part of the announcement, and it lands at a specific tier. OpenAI’s Partner Network opened on June 14, 2026 with a stated $150 million behind it and three membership levels: Select, Advanced and Elite. OpenAI says each level sets a bar on sales performance, technical capability, co-selling with its own teams, and deployment experience. EPAM enters at the middle level. It was not in the launch group, which included Accenture, Bain, BCG, McKinsey’s QuantumBlack unit, PwC and Eliza, and where KPMG later reached the top Elite tier.
What the certification commitment covers
The 5,000 figure is pointed at one job title. EPAM’s pitch to buyers rests on forward-deployed engineers, staff embedded at the customer’s site rather than working from a delivery center, and the company says the training program for them will be built jointly with OpenAI. OpenAI is running an adjacent pilot of its own, a Forward Deployed Experts program with a set of founding partners that pairs partner practitioners with OpenAI’s internal forward-deployed engineering teams.
Against EPAM’s own payroll the target is meaningful. The company reported roughly 56,500 delivery professionals as of March 31, 2026 in its first-quarter results, so certifying 5,000 consultants in a year touches close to one in eleven of them. Against OpenAI’s program-wide goal of 300,000 certified consultants by the end of 2026, it is one contributor among many.
EPAM is now carrying two of these programs simultaneously. In May 2026 it announced a multi-year partnership with Anthropic built around 10,000 Claude-certified architects, including 250 forward-deployed specialists it calls Black Belts, with more than 1,300 certified at the time. The pattern is now standard across the large integrators: Cognizant reached the top tier of Anthropic’s partner program a day before EPAM’s OpenAI designation, on the strength of a comparable certification build-out.
What is already running
EPAM says the OpenAI work is in production in at least one named account. 1&1, the German telecommunications provider, built an agentic customer-service platform with EPAM that now handles a significant share of its weekly call volume through more than 20 AI agents, with the first production deployment completed in under three months. Those figures come from EPAM.
The practice is organized around two tracks, according to the companies’ joint partner page. One puts OpenAI models inside the software development lifecycle, including scanning and remediating vulnerabilities at the point of build. The other re-engineers business processes in customer service, marketing, sales and operations. That second track is where OpenAI’s own research is pointing: the company has found employees reaching for ChatGPT well outside their formal job descriptions, and carrying that fluency into governed, regulated systems is the gap EPAM says its engineers close. In a companion post, EPAM described the same foundation applying across its model vendors: cloud, on-premises and hybrid deployment, an AI-assisted development lifecycle, and security work embedded through it.
Colleen Kapase, OpenAI’s vice president of strategic global partners and ecosystems, is quoted in the release welcoming EPAM at the Advanced level and pointing to its engineering and enterprise-transformation record. Elaina Shekhter, EPAM’s senior vice president and chief strategy and transformation officer, framed the work as turning frontier models into production outcomes rather than deployments for their own sake.
Why the timing matters
The designation arrives with EPAM’s growth guided modestly. Revenue rose 7.6% year over year to $1.400 billion in the first quarter of 2026, and 3.7% on an organic constant-currency basis. For the full year the company guided to revenue growth of 4.0% to 6.5%, with second-quarter revenue of $1.400 billion to $1.415 billion, or 4.0% growth at the midpoint. AI-native work is the line management points to for acceleration, and a vendor certification program is the mechanism by which a services firm converts a model maker’s roadmap into billable scope: certified headcount, co-sell motions, and named production references.
EPAM reports second-quarter results on August 6, 2026, the first set of numbers to arrive with the OpenAI practice in place.