Data Engineer Payments & Funds-Flow
Posted 5
views
Summary
Data engineer owning the payments and funds-flow data domain for OSL/Banxa, a regulated crypto payments business in Singapore. Day to day: modelling messy multi-source payment flows into reconcilable, audit-ready datasets on MaxCompute/DataWorks with CDC ingestion, setting data-quality SLAs, and using AI agents/LLMs as a core part of building and operating data.
About the Role
OSL Group is a listed digital-asset and fintech group; Banxa is our regulated crypto payments and fiat on/off-ramp business. This role is the data owner for the payments and funds-flow domain, serving the payments, finance, compliance and risk teams. Core stack is MaxCompute and DataWorks with CDC-based ingestion. The role works in an AI-native way by default.
What We're Really Looking For We care more about how you think and the leverage you create than about any specific tool. Modelling judgement.
You turn messy, multi-source payment flows into clean, reusable models, and can defend your choices on grain, dimensions and trade-offs. You've owned at least one subject area end to end. Correctness under compliance.
You treat financial data as something that must reconcile and survive audit — idempotent, precise, verifiable — not just 'loaded'. Ownership over ambiguity.
You engage upstream at product kick-off, pin down definitions before writing code, and stay accountable from ingestion to the business outcome rather than stopping at your layer. Influence without authority.
You get backend, SRE, finance and compliance teams to move with you, and you set the quality bar and SLAs for a domain rather than following someone else's. AI-native leverage.
You use AI agents/LLMs as a core part of how you build, review and operate data — automating the repetitive, accelerating the hard, and building data agents and copilots that raise the whole team's throughput. You have a point of view on where AI helps and where it must not be trusted. Must-Have 8+ years in data engineering or data warehousing, including 2+ years in payments, settlement, financial services or exchange environments. Strong SQL and dimensional modelling, with end-to-end ownership of at least one subject area. Production CDC ingestion experience, including late and out-of-order events, deduplication, soft deletes and schema evolution. Accuracy engineering for financial data: idempotent writes, numeric precision, reconciliation and cross-system consistency checks. Agent-based development capability: multi-step agent workflows, tool and data-source integration through MCP, and reusable skills; context engineering, tool-call design, agent evaluation and debugging. Ownership of a data quality framework, covering rule design, threshold rationale, alert response and SLA commitments. Preferred Crypto or digital-asset data experience: on/off-ramp, on-chain or stablecoin settlement, multi-asset exposure. Multi-timezone, multi-currency and multi-region data, including business-day cut-off, FX treatment and localisation. Financial audit or regulatory-reporting support, including source-of-truth and explainable change history. Cross-engine migration experience, for example MaxCompute or Hadoop to Snowflake or Databricks. Building metadata, lineage, catalogue, metric platforms — or
AI data agents / LLM-powered tooling . Python or Java for internal tooling and services. Familiarity with compliance constraints on architecture: PCI DSS, GDPR, data localisation, row and column level security, masking. How We Work Definitional rigour
— clarify before you code. Reproducible by default
— idempotent, re-runnable, verifiable. End-to-end ownership
— own the full chain and faces the business directly. AI-native
— highest-leverage path first; automate the routine, keep human judgement where correctness and compliance demand it.
What We're Really Looking For We care more about how you think and the leverage you create than about any specific tool. Modelling judgement.
You turn messy, multi-source payment flows into clean, reusable models, and can defend your choices on grain, dimensions and trade-offs. You've owned at least one subject area end to end. Correctness under compliance.
You treat financial data as something that must reconcile and survive audit — idempotent, precise, verifiable — not just 'loaded'. Ownership over ambiguity.
You engage upstream at product kick-off, pin down definitions before writing code, and stay accountable from ingestion to the business outcome rather than stopping at your layer. Influence without authority.
You get backend, SRE, finance and compliance teams to move with you, and you set the quality bar and SLAs for a domain rather than following someone else's. AI-native leverage.
You use AI agents/LLMs as a core part of how you build, review and operate data — automating the repetitive, accelerating the hard, and building data agents and copilots that raise the whole team's throughput. You have a point of view on where AI helps and where it must not be trusted. Must-Have 8+ years in data engineering or data warehousing, including 2+ years in payments, settlement, financial services or exchange environments. Strong SQL and dimensional modelling, with end-to-end ownership of at least one subject area. Production CDC ingestion experience, including late and out-of-order events, deduplication, soft deletes and schema evolution. Accuracy engineering for financial data: idempotent writes, numeric precision, reconciliation and cross-system consistency checks. Agent-based development capability: multi-step agent workflows, tool and data-source integration through MCP, and reusable skills; context engineering, tool-call design, agent evaluation and debugging. Ownership of a data quality framework, covering rule design, threshold rationale, alert response and SLA commitments. Preferred Crypto or digital-asset data experience: on/off-ramp, on-chain or stablecoin settlement, multi-asset exposure. Multi-timezone, multi-currency and multi-region data, including business-day cut-off, FX treatment and localisation. Financial audit or regulatory-reporting support, including source-of-truth and explainable change history. Cross-engine migration experience, for example MaxCompute or Hadoop to Snowflake or Databricks. Building metadata, lineage, catalogue, metric platforms — or
AI data agents / LLM-powered tooling . Python or Java for internal tooling and services. Familiarity with compliance constraints on architecture: PCI DSS, GDPR, data localisation, row and column level security, masking. How We Work Definitional rigour
— clarify before you code. Reproducible by default
— idempotent, re-runnable, verifiable. End-to-end ownership
— own the full chain and faces the business directly. AI-native
— highest-leverage path first; automate the routine, keep human judgement where correctness and compliance demand it.