ML/AI Engineer — Thunes Salary Negotiation Guide

Negotiation DNA: This guide decodes Thunes' Interoperability mandate, translating the Feb 2026 Swift Direct Global Network link into an ML/AI engineering compensation framework spanning Singapore, London, and New York markets.

---

Compensation Benchmarks (2025-2026)

Region	Base Salary	Options (4yr)	Total Comp
:singapore: Singapore (SGD)	S$130,000 – S$195,000	S$70,000 – S$140,000	S$200,000 – S$335,000
:gb: London (GBP)	£80,000 – £125,000	£45,000 – £85,000	£125,000 – £210,000
:us: New York (USD)	$150,000 – $215,000	$80,000 – $160,000	$230,000 – $375,000

Negotiation DNA: ML/AI Engineers at Thunes build the intelligent systems that power a cross-border payment network spanning 130+ countries and 4B+ mobile wallets. The ML applications here are uniquely valuable: real-time fraud detection across payment methods ranging from SWIFT bank transfers to M-Pesa mobile money; intelligent payment routing that optimizes for cost, speed, and reliability across hundreds of corridors; dynamic FX pricing models; and predictive settlement systems. Thunes is the critical interoperability layer between old-world SWIFT and new-world instant payments, and ML/AI Engineers build the intelligence that makes that bridge smart, not just functional. Your compensation should reflect the direct margin impact of your models.

---

Level Mapping & Internal Benchmarking

Thunes Level	Equivalent at TerraPay	Equivalent at Nium	Equivalent at Wise
ML/AI Engineer	ML Engineer	ML Engineer	ML Engineer
Senior ML/AI Eng	Senior ML Eng	Senior / Lead ML	Senior ML Engineer

Thunes ML/AI Engineers work with more diverse training data than peers at Wise or Nium. While Wise has larger absolute transaction volume, Thunes' data spans more heterogeneous payment systems across more countries. This data diversity produces more robust models and commands a compensation premium.

---

:link: Thunes Interoperability & Standards-Bearer Lever

The Swift Direct Global Network link (Feb 2026) creates extraordinary ML/AI opportunities. By connecting to Swift's infrastructure, Thunes gains access to cross-protocol transaction data that enables new categories of machine learning models.

As a Standards-Bearer ML/AI Engineer, you are building the intelligence layer of the legacy-to-real-time bridge:

• Cross-protocol fraud detection models that correlate patterns across SWIFT's institutional banking transactions and Thunes' digital wallet/mobile money flows
• Intelligent routing models that select the optimal path (SWIFT vs. real-time) for each transaction
• FX optimization models that account for both legacy banking settlement cycles and instant payment pricing
• Demand forecasting models that predict transaction volumes across the interoperability layer

The critical interoperability layer between old-world SWIFT and new-world instant payments generates a unique cross-protocol dataset that no other company can access. Your models will be trained on both legacy banking transaction data and real-time settlement data — this dual-protocol ML capability is the Standards-Bearer premium. Real-time settlement across 130+ countries and 4B+ mobile wallets provides an unparalleled training dataset.

---

Global Levers

Lever 1: Cross-Protocol Fraud Detection

"The Swift Direct Global Network link creates a unique ML opportunity: building fraud detection models that correlate patterns across SWIFT's institutional banking transactions and Thunes' digital wallet flows. No other company has this cross-protocol dataset. I will be building models that detect fraud patterns invisible to single-protocol systems. This is pioneering work that commands premium compensation."

Lever 2: Routing Optimization with Direct Revenue Impact

"The ML models I build determine which payment rail to use for each transaction — SWIFT vs. real-time — optimizing across cost, speed, and reliability. Every basis point of routing optimization at Thunes' transaction volume translates to millions in margin. This is ML with direct, measurable commercial impact."

Lever 3: Unique Training Data Advantage

"Thunes' cross-border transaction data spanning 130+ countries and multiple payment protocols is one of the most unique ML training datasets in fintech. The models I build here leverage data that cannot be replicated elsewhere, which makes both my work and the resulting IP exceptionally valuable."

Lever 4: Pre-IPO AI Asset Value

"ML models and training data are increasingly valued as intellectual property in payments company valuations. The fraud detection, routing optimization, and pricing models I build become a durable competitive advantage. I would like an options package that reflects my contribution to building this long-term IP moat."

---

Negotiate Up Strategy: In Singapore, start at S$190,000 base with S$130,000 in options (4yr vesting) and accept no lower than S$150,000 base with S$80,000 in options. In London, start at £120,000 base with £80,000 in options and accept no lower than £90,000 base with £50,000 in options. In New York, start at $210,000 base with $150,000 in options and accept no lower than $165,000 base with $90,000 in options. Lead with the cross-protocol fraud detection and routing optimization angles — you are building ML models that directly drive margin on a multi-billion-dollar payment network.

---

Evidence & Sources

• Thunes company profile, ML applications, and network data (thunes.com)
• Swift Direct Global Network cross-protocol data integration (Feb 2026)
• Glassdoor and Levels.fyi ML/AI Engineer compensation data for Singapore, London, NYC fintech (2025)
• Wise, Nium, and TerraPay ML team structures and compensation benchmarks
• Academic research on cross-border payment fraud detection and routing optimization
• Industry analysis of ML/AI IP valuation in fintech (PitchBook, CB Insights)