Senior Backend Engineer, Payments

About the role

Junction Pay moves money for roughly 9,000 Hong Kong merchants, from Sham Shui Po electronics stalls to cross-border e-commerce sellers shipping into the Greater Bay Area. Payments Platform Engineering owns everything after the button press: authorisation, the double-entry ledger, settlement batches, refunds and the reconciliation jobs that have to balance to the cent before Finance opens the books each morning.

We are eighteen months into moving the ledger off a single Postgres primary and onto a partitioned, event-sourced design. The migration runs alongside live traffic — nothing gets a maintenance window — so the work is equal parts distributed-systems design and careful, boring, reversible change. This role exists because the two engineers who started that migration are now leading their own squads and we need a third pair of senior hands on the core.

You would join a squad of six (four engineers, a product manager and an SRE) sitting in Cyberport, with a fortnightly cadence of design reviews that anyone in the company can attend and heckle. Expect to spend real time with the Financial Crime and Regulatory Affairs teams — in a licensed stored-value business, engineering decisions become licence conditions surprisingly fast.

What you'll do

• Design and ship services in Go against Postgres and Kafka, with an appetite for idempotency, exactly-once semantics and the failure modes in between
• Extend the double-entry ledger to support multi-currency settlement without breaking the existing reconciliation guarantees
• Cut over legacy authorisation paths to the new event-sourced core using dual-write, shadow-read and staged traffic shifting
• Own production for what you build: on-call rotation is one week in six, with a hard rule that pages you can predict get fixed rather than documented
• Write the technical sections of change requests that Regulatory Affairs files with the regulator, and defend them in review
• Mentor two mid-level engineers through design docs and code review — we measure senior engineers on the team's output, not their own

What we're looking for

• 6+ years building backend services in production, at least 2 of them on systems handling money, orders or another domain where a lost write is unacceptable
• Strong Go, or strong Java/Kotlin/Rust plus a genuine willingness to write Go from week one
• Deep working knowledge of relational databases — transaction isolation, locking behaviour, partitioning and query plans, not just the ORM
• Practical experience with an event log (Kafka, Pulsar or similar) and the operational reality of consumer lag and replay
• Comfortable operating in English across engineering, product and compliance; this squad works in English end to end

Nice to have

• Exposure to card scheme messaging, ISO 20022 or a local faster-payment scheme
• Experience supporting an audit or regulatory inspection of a system you built
• Terraform and Kubernetes fluency — our platform team runs it, but you will read it

What you get

• Hybrid: two fixed days at the Cyberport office, the rest as you like
• Medical and dental for you and dependants from day one, no waiting period
• 20 days annual leave rising to 25, plus the office closes between Christmas and New Year
• HK$12,000 a year of learning budget, historically spent on conferences rather than courses
• Visa sponsorship and relocation support for the right candidate