Sixty-seven percent of finance AI engineers in the ENTRA Q3 2026 Candidate Survey are actively benchmarking roles outside traditional finance on Labor Day week. JPMorgan Chase and Goldman Sachs enforce five-day mandates with badge-swipe data; Ramp, Brex, and Stripe NYC run two-to-three day hybrid schedules with no enforcement mechanism. The engineers caught between these two postures are doing the math, and the math points toward the door.
The 5-Day Mandate: JPMorgan and Goldman's Enforcement Turn
JPMorgan Chase moved to full five-day RTO for all staff in Q1 2024 — no role-based exceptions, no hybrid carve-outs. By Q3 2026, the policy has entered its enforcement phase. Badge-swipe data is now used systematically to identify non-compliance, and the message delivered through management chains across the firm's technology and AI organizations is unambiguous: presence is the expectation and absence is documented.
Goldman Sachs implemented a matching five-day standard with equivalent language — "no exceptions" — but the practice inside Goldman's AI research function has diverged from the policy text. Senior AI researchers and quant model developers have, in practice, been granted informal flexibility that does not appear in any published policy document. Two Goldman AI engineers who spoke on condition of anonymity described the arrangement as "exceptions that aren't called exceptions." The asymmetry matters: junior AI engineers see the written rule and plan around it; senior talent negotiates informally and often wins flexibility the firm will not admit exists publicly.
BlackRock's posture is the most legible bifurcation on the Street. Research staff — including the AI and quantitative modeling teams embedded within BlackRock's Aladdin platform — operate on a four-day hybrid schedule. Technology and AI engineering staff, a separate classification in BlackRock's HR taxonomy, are on a five-day mandate. The line between "research" and "engineering" in a modern machine learning context is increasingly thin — a mid-level MLE building Aladdin's risk model inference layer is doing work that in any frontier lab would be classified as research — but BlackRock's policy draws the boundary at the team designation, not the job content.
The enforcement data tells a specific story about NYC Local Law 32 compliance alongside RTO posture. JPMorgan's salary band posting compliance rate is 89%, per ENTRA Q3 2026 tracking of NYC-based job listings. Goldman Sachs posts salary bands on 76% of qualifying NYC postings. Both figures represent improvement from earlier in 2026, but both sit below the rates posted by their fintech counterparts. The compliance gap is not incidental — it correlates with the broader transparency posture each institution holds toward the external talent market.
The Hybrid Counterpoint: Ramp, Brex, and Stripe NYC
Ramp and Brex each post salary bands on 100% of their NYC-listed roles. Stripe's New York operation matches that rate. The transparency is not accidental — it is a deliberate signal to a candidate pool that has grown acutely sensitive to information asymmetry after years of negotiating with banks that disclose ranges reluctantly.
Ramp operates a two-to-three day hybrid schedule from its Manhattan office, with no badge-based enforcement. Brex, which maintains a significant New York engineering presence alongside its San Francisco operations, runs the same schedule with explicit policy language framing flexibility as a design feature rather than a concession. Stripe NYC — part of the San Francisco company's distributed engineering structure — holds to the Stripe-wide model: teams coordinate their in-office rhythm; the company does not mandate it from above.
The organizational logic behind these postures is straightforward. Ramp, Brex, and Stripe are competing for engineers who can work at any AI lab or major technology company. Their compensation packages — $175,000 to $255,000 base for AI engineering roles, per ENTRA Q3 2026 Candidate Survey data — sit below the floor of what JPMorgan's AI engineering bands pay at comparable seniority levels. The fintech-native companies cannot win on base compensation. They compete on equity upside, mission framing, and work arrangement — and they have made a deliberate choice to lead with the last of those three in a market where RTO friction at traditional finance firms is at its highest recorded level.
The 100% salary posting compliance rate is part of the same strategy. When Ramp lists an AI engineering role with a published band of $185,000 to $220,000 base in New York, it is signaling to every Goldman AI engineer operating under a 76%-compliant posting regime that the information relationship between employer and candidate will be different at Ramp. That signal is worth more in a Labor Day week characterized by a 34% spike in AI engineer job searches — the ENTRA Job Signal Index Q3 2026 reading for the week ending September 6 — than any individual role posting.
The Talent Math: Comp Differential, Exit Signals, and the Labor Day Spike
The compensation picture is more complex than the "startups pay less" framing that circulates in finance HR discussions.
Finance AI engineering at JPMorgan and Goldman runs $195,000 to $285,000 in base compensation for senior individual contributor roles, per ENTRA Q3 2026 Candidate Survey aggregates from 412 NYC-based finance AI candidates. Startup AI at Ramp, Brex, and Stripe NYC runs $175,000 to $255,000 in base — a discount of roughly $20,000 to $30,000 at comparable seniority levels. On base alone, traditional finance holds the advantage.
The equity structure inverts the picture. Finance AI engineers at JPMorgan and Goldman hold RSUs tied to institutions with market capitalizations in the hundreds of billions — stable, liquid, and with limited upside relative to current price. Ramp's most recent valuation stands at $13 billion following its Series D-2; engineers who joined during its Series C at a $5.8 billion valuation have already seen meaningful appreciation. Brex, which reached a $12 billion valuation before the 2022 fintech correction and has since restructured its equity program for new hires, offers a more complex equity story — but its current-cycle new hire packages carry 2x to 4x the paper upside of a JPMorgan RSU package at comparable base, per ENTRA Q3 2026 comp survey data.
The 67% exit-benchmarking rate in the ENTRA Q3 2026 Candidate Survey is not a sentiment reading — it reflects active behavior: candidates who have applied to at least one non-finance AI role or accepted a recruiter conversation in the 30 days prior to survey completion. The Labor Day week search spike to 34% above baseline in the ENTRA Job Signal Index Q3 2026 data suggests that the benchmarking converts to formal job search activity at a higher rate in September than in any other quarter, consistent with the annual "back to work" decision window that senior engineers use to evaluate whether another year in their current role is the right call.
The specific roles generating the most outbound mobility signal from finance AI are: ML infrastructure engineers at the L6/VP equivalent level at JPMorgan, quantitative model developers at Goldman with NLP and LLM fine-tuning backgrounds, and data science leads inside BlackRock's Aladdin engineering group. These are not generalist finance technologists — they are engineers with skill sets that Ramp's risk modeling team, Brex's fraud detection infrastructure, and Stripe's payments ML stack actively recruit for. The talent pool is small, the overlap is near-total, and the RTO differential is the single most frequently cited variable in ENTRA candidate conversations with this cohort.
What September Brings for NYC AI Talent
Three signals define the next 90 days in New York's finance AI talent market.
JPMorgan's badge-swipe data becomes a retention instrument. Systematic enforcement of five-day attendance creates a documented record of compliance. Managers who know their attrition risk candidates are using that data are more likely to grant informal flexibility that mirrors what Goldman's AI research team quietly operates — but JPMorgan's stated policy offers less cover for such arrangements. Expect the informal exception culture that exists at Goldman to develop inside JPMorgan's AI org through Q4, or for a further round of senior ML exits that becomes visible in LinkedIn departure signals by November.
Ramp's and Brex's next hiring rounds will land differently. Both companies have funding runway and open AI engineering headcount. In a Q3 market where 34% more AI engineers are searching than baseline, and where 67% of finance AI candidates are benchmarking externally, the timing of their next public hiring waves will capture a candidate cohort that has done its comparison shopping. The firms that post first — with full salary bands and explicit hybrid policy language in every job description — will convert at higher rates than in any prior quarter.
NYC Local Law 32 enforcement pressure intensifies. The New York City Commission on Human Rights has begun issuing cure letters to employers with systematic non-compliance patterns in 2026. At 76% compliance, Goldman is the largest finance employer in New York running material non-compliance risk. A formal enforcement action against a major bank over salary posting requirements would shift the NYC AI talent market's information environment overnight — and would do more to close the transparency gap between traditional finance and fintech-native employers than any voluntary policy update.
Labor Day 2026 is not the beginning of the RTO war in New York finance. It is the point at which the casualty count becomes legible.
Global Career PlatformFind AI talent. Find your next role.
Booking is hotels. · Airbnb is apartments. · ENTRA is global careers.