Every dollar that flows through Deriv's payment infrastructure needs to end up in the right place. Your job is to make sure it does. We process payments through 53+ payment service providers across fiat and crypto. Every day, deposits, withdrawals, fees, reserves, and settlements have to reconcile across three systems: the PSP, our back office, and the ERP. When the numbers don't match, you're the person who finds the gap and figures out why.
Why This Matters
This isn't a month-end-only role. The team works on a daily reconciliation tempo. Discrepancies are flagged, investigated, and resolved within hours, not weeks. You're operationally close to the data, technically curious about why things break, and direct in escalating issues to the right people.
Deriv is mid-build on an AI Reconciliation platform that will eventually run most of this work autonomously. The platform is already in production and getting smarter every release, but it depends on clean upstream data and human judgment for the edge cases. You're the person who knows each PSP's quirks, spots when something doesn't look right, and helps the platform separate signal from noise.
The recon team is also the last line of defence. When something slips through upstream systems, you're the one who catches it. We didn't notice isn't an acceptable answer for the team you'll be joining.
Why Deriv
Our finance team is already shipping production AI. Finance automation covers invoice processing, reconciliation, and real-time anomaly detection. Not prototypes. Production systems running continuously across 2.5M traders and 53+ payment providers.
You'll join a team that's building the tools you'll use. The AI Reconciliation platform you'll be testing and shaping is the same one that will eventually handle the work at scale. Your feedback directly influences what gets built next.
Real money, real regulations, real consequences. If that sounds like the kind of environment where you do your best work, we'd like to hear from you.
What You'll Do
Daily PSP Reconciliation
- Run daily reconciliations across 53+ fiat and crypto PSPs: deposits, withdrawals, fees, reserves, settlements, opening and closing balances
- Investigate discrepancies, classify them (data issue, currency mismatch, timing gap, real loss), and drive resolution with PSPs, engineering, or Backoffice as appropriate
- Maintain the three-way match: PSP source data, Backoffice records, ERP system
PSP Data Quality
- Run daily checks on browser-automation and API-based PSP data feeds, flagging stale data, IP blocks, dataset format changes, and missing currencies
- Handle manual data uploads when automation breaks, with clear escalation to engineering for permanent fixes
- Own the working knowledge of each PSP's quirks: which ones change formats often, which ones swap deposit and withdrawal labels, which ones report in IDR or KES or UGX, and what normal looks like for each
AI Recon Platform Partnership
- Run UAT testing on new platform features and fixes, validating that recon logic works as intended on real production data
- Raise and prioritise bugs with the AI engineering team, including detailed reproduction examples (transaction IDs, sample data, expected-vs-actual)
- Help shape platform improvements by surfacing naming conventions, transaction-type labelling, and exception-handling logic that makes the platform smarter
Month-End Closing Support
- Run month-end PSP balance reconciliation, including chasing PSP confirmations where needed
- Document unresolved exceptions and ensure they don't carry forward silently
- Support the broader finance team on revenue, settlements, and crypto accounting reconciliations as needed
Process Documentation
- Maintain process documentation for PSP data extraction and recon workflows so the team isn't dependent on tribal knowledge
- Onboard the team on new PSPs as they're added to the portfolio: the data sources, recon logic, and quirks specific to that provider
Experience
Who You Are
- 3+ years in reconciliations, accounts operations, or finance operations. Payments, fintech, or trading platform experience is a plus, not a requirement
- You've reconciled real third-party data before, not just internal GL movements. Whether it's PSPs, banks, or any other external counterparty at meaningful volume, you know what a settlement file looks like, what doesn't tie out, and how to chase it down
Technical Skills
- Comfortable with data at scale. You can write or read enough SQL or BigQuery to investigate a discrepancy without waiting for an engineer. Excel and Sheets are second nature
- Solid grasp of accounting fundamentals. You don't need to recite IFRS, but you understand debits and credits, opening and closing balances, accruals vs cash, and FX impact well enough to apply them on a live reconciliation
- Working understanding of payment processing: what authorisation, capture, settlement, chargebacks, fees, and reserves mean in a PSP statement
- Curious about crypto, not afraid of it. You don't need wallet experience, but you should understand the basics: what crypto is, how to verify a transaction has been received on-chain, and why USDT on different chains isn't the same thing
How You Work
- Sharp pattern-recognition. You notice when daily volumes drop unexpectedly, when a currency that should be there isn't, when a recon passes but feels wrong
- Direct communication style. You raise issues clearly in Slack, attach examples, tag the right people, and don't sit on problems hoping they'll go away
- You like AI tools and use them to do your job better: pulling logic apart, drafting Slack messages, summarising long threads, automating repetitive checks. We're an AI-first finance team and you'll fit in with people who experiment
The Honest Reality
This isn't a role where you run a fixed monthly checklist and clock out. Discrepancies happen daily. PSPs break their data formats with no warning. The AI platform isn't perfect yet and sometimes you'll spot bugs the engineering team missed. Some days are operationally heavy; some days are about driving longer-term fixes with engineering.
You'll inherit a mix of mature processes and things still being figured out. The pace is fast. The standard is high. But the team is supportive, the engineering partnership is real, and the AI tooling we're building is genuinely meant to make your job better, not replace your judgment.
If you want quiet, predictable work with no ambiguity, this isn't it. If you want to be on the team that catches what others miss and helps shape the AI tools that finance teams will be using five years from now, it might be.
