How EPAM Systems Pays

Technical ladder — L1 through L9

Management ladder — M1 through M5

Disclosed executive layer

Track divergence

Hierarchy qualifications and legacy structures

• The L1 to L9 sequence is reconstructed from employee-reported salary submissions, not from an EPAM document. Every level in it is labelled Reconstructed in the source report, including the four levels that carry directly reported medians.
• Cross-country title equivalence is imperfect. Lead, Chief Engineer, Solution Architect and Delivery Manager can carry very different scope depending on account size, client-site status and the local labour market, so the same title is not the same job in Kraków, Bangalore and Newtown.
• The management tiers below director are priced at their stated technical equivalents because EPAM discloses no management-track pay at all. That makes M1 and L4 identical in the money columns even though the experience windows differ.
• Only L1 through L4 have directly observed medians in most markets. L5 through L9 are extrapolated from L4 using fixed 1.18x, 1.38x, 1.62x, 1.92x and 2.25x multipliers, which means the shape of the senior ladder is an assumption applied identically to every country.
• Equity below the named executives is a scenario band, not an entitlement. EPAM's plan permits awards to employees, consultants and service providers, but public data shows no routine grant below senior levels and no minimum eligibility band anywhere.
• The median employee EPAM identified for its pay ratio was a full-time salaried professional in Eastern Europe on $38,570, which tells you where the mass of the ladder actually sits: roughly 95% of the eligible population is outside the United States.

Peer-level mapping

Office and market catalogue — calibration factors versus Newtown, Pennsylvania

Presence for all 42 markets comes from EPAM's official global hiring-location pages. Every market can contain both delivery and client-facing work; the profile labels describe the dominant mix the source report observed, not an exclusive one. The United States, United Kingdom, Australia, Switzerland, Singapore, Canada and the UAE skew client-facing and onshore, while India, Ukraine, Poland, Hungary, Georgia, Armenia, Kazakhstan, Uzbekistan, Mexico, Colombia and Spain skew delivery.

Newtown, Pennsylvania anchor medians — the basis of every modeled cell

• Newtown, Pennsylvania is EPAM's global headquarters and the anchor city at a factor of 1.00. It is also the company's largest H-1B worksite, with 166 filing positions at a $115,000 certified median.
• India was the largest delivery location at roughly 12,200 professionals at December 31, 2025, up about 2,150 year over year, followed by Poland at roughly 5,050, Belarus at roughly 3,400 and Mexico at roughly 2,950.
• The city table deliberately contains no Russian location and no Belarusian location. EPAM's current hiring-location page evidences the post-2022 footprint, in which delivery capacity shifted toward India, Central and Western Asia, Latin America and Central and Eastern Europe while Belarus continued to shrink in relative terms.
• Ukraine remains a strategically significant delivery market rather than a legacy one. Kyiv, Lviv, Kharkiv and Dnipro all remain in the model, with Lviv carrying the country's reported anchor.
• New Zealand appears in the source bundle's component schema only because the requested template called for it. EPAM's current global hiring-location page did not verify a New Zealand presence on the research date, so Auckland is excluded from this page entirely.
• Switzerland is the only market in the model that sits above the US anchor, at about 1.14x of Newtown at lead level and above. Every other market sits below it.

Executive incentive targets — percent of salary

Named-executive outcomes

• Stockholders added 4.0 million shares to the 2025 LTIP on May 21, 2026, taking the stated reserve from 2.5 million to 6.5 million shares plus returning shares. The ESPP gained 650,000 shares in the same meeting, approved with 45,718,624 votes for, 308,121 against and 56,864 abstentions.
• Before the amendment, 512,838 shares remained of the original 2.5 million reserve, about 20.5% of the stated pool. A precise post-amendment utilisation rate cannot be calculated from public data because grants between March 16 and May 21 and the returning-share mechanics both move the denominator.
• Outstanding options carried a $213.60 weighted-average exercise price at March 16, 2026 against a $110.25 reference share price in August 2026, so the legacy option overhang was deep underwater. That is the practical reason the 2025 named-executive mix contains no new options at all.
• During 2025 EPAM issued approximately 213,000 shares through the ESPP. The option roll-forward records 438,000 exercises while the stockholders' equity statement shows 417,000 shares issued from exercises; the 21,000-share difference reflects different presentation and measurement contexts in the filing and should not be reconciled without the award-level ledger.

Plan reserves and outstanding awards at March 16, 2026

Vesting — common reported employee schedule

Time-based RSUs vest 25% annually over four years with no cliff beyond the first anniversary, disclosed for named executives and corroborated by employee-reported vesting data. Balazs Fejes's March 15, 2026 grant of 26,238 units is scheduled to vest 25% annually from March 2027 through March 2030. Executive PSUs instead use a three-year performance period weighted 37.5% adjusted revenue growth, 37.5% adjusted EPS and 25% relative total shareholder return, paying 50% to 200% of target when thresholds are met. Director RSUs vested after one year for the 2025 annual award.

Eligibility by hierarchy level

• The plan permits awards to employees, consultants and service providers, but EPAM discloses no minimum eligible band, no country availability table by level and no standard grant frequency. Equity below senior levels is a possibility, not an entitlement.
• Public data shows no recurring grant at L1 or L2. From L3 the source report describes selective or discretionary grants, and from L7 upward a high probability of critical-talent eligibility, but every grant value in the band table is a scenario midpoint rather than a disclosed figure.
• The ESPP is the clearest broad-based ownership route and it works differently from a free grant: it is payroll-funded, available in select countries only, and delivers its value through a 15% discount on the lower of two prices rather than through a grant.
• The 2025 named-executive mix was 50% PSUs and 50% time-based RSUs with no new options. That eliminates option leverage for the top of the house while increasing full-value share dilution and retention value.
• Non-employee directors each received 2,132 RSUs on May 21, 2026, against 1,372 RSUs on May 22, 2025 with a $247,824 grant-date fair value. Employee directors Fejes and Dobkin receive no additional director compensation.

Indicative annual grant value by band — Newtown, Pennsylvania

Annualized planning value (±25%), not the face value of every new-hire grant. Share equivalents use the $110.25 reference price at 2026-08-26 and ignore plan valuation rules and PSU performance. Highlighted rows are disclosed grant-date stock awards.

Named executive target equity

Summary Compensation Table values are grant-date accounting values, not cash realised during 2025. The most striking feature of the roster is that the founder-Executive Chair out-earned the sitting CEO on every line: an $850,000 salary against $756,300, a $1.08M cash incentive against $767,000 and $6.19M of stock against $3.72M. Dobkin also holds the deep legacy option position, exercising 41,026 options at a $71 strike in March 2026, and adopted a Rule 10b5-1 plan on March 13, 2026 permitting sales of up to 150,000 shares through March 15, 2028.

Board compensation framework

Nine non-employee directors were paid between $322,100 and $406,800 for 2025, and the spread between them comes almost entirely from committee and role retainers rather than from the equity award, which was identical for everyone at 1,372 RSUs worth $247,824. Board equity is therefore roughly 61% to 77% of a director's total, and it comes from a separate 487,007-share pool that does not compete with the employee LTIP.

Not publicly disclosed

Modeled promotion planning timeline

The modelled promotion clock runs 1-2 years from L1 to L2, 2-3 years from L2 to L3, then 2-4 years for each step from L3 through L7 and 3-5 years for L7 to L8 and L8 to L9. Reaching Distinguished Engineer therefore takes roughly 18 to 30 years of consistent promotion, which matches the 18+ years of experience the level itself carries. Every one of those intervals is modelled: skill certification, project scope and whether an open role exists at the next level all move it.

Dataset summary

City-level H-1B wage history

All-years aggregates. P25/P75 are rounded reconstruction values marked modeled; medians and record counts are the stronger fields.

Selected title / worksite records

Evidence classification

Explicit “Not publicly disclosed” index

Known gaps and diligence before relying on a cell

1. 1Base and total calibration factors are identical in every market because the source bundle's total-compensation figures are base plus a variable target, and its equity numbers are one global USD scenario band applied to every country. Setting the total factor from anything other than the base-pay calibration would have manufactured a drift toward 1.00 offshore that the evidence does not support.
2. 2Anchor totals were rebuilt from base plus bonus plus equity so that every row reconciles. That changes the senior-level totals against the source report, whose total-compensation figures exclude equity entirely: the L9 anchor reads $531,250 here against $357,750 there, with the difference being the $200,000 midpoint of the bundle's own equity scenario band.
3. 3The geographic interpolation is linear in seniority, but the evidence is not. Observed ratios rise steeply from L1 to L4 and then stop, because every level above L4 was extrapolated from L4 with identical multipliers in every country. Seniority spacing was compressed at the top of the ladder so both endpoints land exactly, which leaves a mid-ladder residual: the model understates the L2 and L3 factors in India by roughly a fifth to a quarter, and overstates the Ukrainian L2 factor by a similar margin in the other direction. Where a precise mid-ladder figure matters, use the reported country anchors rather than the interpolated cell.
4. 4Executive rows are carried in the executive section only and are deliberately absent from the band table, so that a single global CEO or NEO figure cannot drag every location's factor toward 1.00 at the top of the ladder. The most senior band is the modelled Vice President tier, which is excluded from the range chart because its $900,000 equity midpoint dwarfs every employee level.
5. 5Equity is scaled by the location factor along with cash. EPAM's LTIP is share-denominated, so an offshore employee who actually receives a grant would get closer to the full USD amount than the scaled figure shows. The scaled number is a total-reward planning envelope, not a grant estimate.
6. 6No employer benefit load, payroll tax or statutory contribution is included in any total. Provident fund, superannuation, social insurance and the 401(k) match are employer cost, not pay, and folding them in would inflate offshore totals most.
7. 7The management ladder is priced at the technical equivalences the source report itself states, so M1 through M3 carry the same money as L4 through L6. EPAM publishes no management-track pay, and the director and VP rows use scenario midpoints for both variable pay and equity.
8. 8The US L1 anchor of $123,000 sits above the L2 anchor of $115,100, and Switzerland's reported L3 exceeds its L4. Both are small-sample artefacts preserved from the source rather than smoothed, and both distort the entry-level factors that anchor the interpolation.
9. 9Auckland appears in the source bundle's component schema only because its template called for New Zealand coverage. EPAM's current global hiring-location page did not verify a New Zealand presence on the research date, so it is excluded from this page entirely.
10. 10Aggregators disagree on FY2025 H-1B filings, showing 122 on one and 101 on another, because of different fiscal cut-offs and inclusion rules. The FY2026 figure of 79 comes from a single snapshot dated August 18, 2026.
11. 11The 2025 share count for option exercises differs between the option roll-forward at 438,000 and the stockholders' equity statement at 417,000. That 21,000-share gap reflects different presentation contexts in the filing and is not reconciled here.

FX rates used — 1 USD equals, snapshot 2026-08-26

Single-date conversion layer for comparability; it ignores payroll-date FX, tax, purchasing power, benefits valuation and hedging. Compensation intelligence, not legal, tax, investment, immigration or employment advice.

Source register — 13 sources

EPAM added about 8,050 people in 2024 and only about 1,650 in 2025, so growth slowed sharply while utilisation kept climbing. Attrition fell to 8.5% in the same period. A firm hiring less, retaining more and running its bench harder is not a firm under pressure to bid up salaries, which is the context for every negotiation figure on this page.