I'm a freelance senior backend engineer working with brokerages and fintech products. Ledgers, KYC, clearing and payments — the parts that have to be right the first time.
Broker-side and payment backends: the account, the money, the integrations and the financial logic underneath a product.
Double-entry ledgers, balances, holds, fee engines and end-of-day reconciliation across equities and options.
Identity and document verification flows, account lifecycle, sanctions screening, audit trails that stand up to review.
ACH deposits and withdrawals, returns and reversals, banking and clearing integrations, settlement timing.
Services on AWS, defined in Terraform and shipped through CI/CD, with observability and least-privilege access wired in. Boring on purpose.
Founding backend engineer on a US regulated equities & options brokerage (2023 — now, client under NDA). Built the core as the sole engineer for the first two years; now own the backend across a platform of roughly 40 deployable units — a core API, ~16 scheduled jobs and lambdas, workers and integrations — all on AWS.
Deposits, withdrawals and transfers with holds and settlement timing; balances, transactions and PnL kept consistent with the clearing firm's books and reconciled daily.
Two-way sync with the clearing firm: account opening, balances, positions and transactions pulled on schedule, webhooks consumed in real time, with replay tooling for when the other side has a bad day.
Application flow from waitlist to funded account: identity and document verification, sanctions screening, account state machine, and JWT auth validated at the API edge.
Trader onboarding into the broker's OMS and risk manager, platform entitlement sync for several trading front-ends, commission recalculation, and the CRM/support bridge ops actually use.
Services talk over SNS/SQS with shared message contracts; idempotent consumers, dead-letter queues and alarms so a missed event is a page, not a silent gap in someone's balance.
I write the Terraform for every service, job and lambda I ship — ECS task definitions, queues, schedulers, IAM, secrets, alarms, CI/CD — on top of shared modules maintained together with DevOps.
Long-term, remote, as a registered independent professional in Spain. I work EU hours and overlap most of the US business day, so US teams get me live, not async. These are the things a client keeps me around for.
Every balance change is a ledger entry. Every external call is idempotent. If a reconciliation can drift, it will — so it's checked daily, not quarterly.
I take systems end to end: data model, integrations, infrastructure, and the admin tooling the ops team actually uses. No handoffs into the void.
Audit trails, access controls and reporting are designed in from day one, because retrofitting them into a live financial system is the expensive version.
Deliberately conventional. The interesting problems are in the domain, not the tooling.