Senior software engineer

Distributed systems, payments infrastructure, and financial platforms.

I specialize in billing, ledgers, payouts, and event-driven backends—building with Java, Kotlin, TypeScript, Postgres, Kafka, Redis, and AWS. I focus on systems that stay reliable and maintainable at scale.

View selected work Email me Download résumé

Selected engineering work

Systems where money, scale, and reliability meet

Billing, ledgers, payouts, and event-driven platforms built for high-volume production systems.

AirbnbSubscription Billing Platform

Recurring subscriptions, invoicing, and payment collection

Worked on backend systems responsible for recurring subscriptions, monthly invoice generation, and payment collection. The system had heavy load spikes at the beginning of billing cycles, so we moved invoice generation and payment attempts from synchronous request paths into asynchronous Kafka workflows to smooth traffic and protect the primary Postgres database.

A major focus was reliability: payment retries, dunning logic, and idempotent charge requests were implemented to ensure customers were never double-charged during retry scenarios or downstream payment processor failures. We also performed Postgres index migrations and historical backfills to support new billing queries without locking large tables during peak hours.

Technologies

Kafka
PostgreSQL
Redis
Java
Kotlin

Concepts

Idempotency
Async workflows
Billing cycles
Retries
Dunning
Observability

AirbnbFinancial Ledger & Reconciliation

Ledger system for payments, refunds, and payouts

Built and maintained a financial ledger service used to record billing events, payments, refunds, and payouts. The ledger was designed as an append-only system to maintain a full audit trail and support reconciliation with external payment processors and bank reports.

After a production incident caused by duplicate ledger entries during a retry storm from a payment processor, we introduced idempotent write paths and built reconciliation jobs that compared internal ledger data against processor settlement reports and flagged discrepancies for finance. This system was used by finance and operations teams for reconciliation, reporting, and audits, so correctness and traceability were more important than latency.

Technologies

PostgreSQL
Kafka
Java

Concepts

Double-entry ledger
Reconciliation
Audit trail
Idempotent writes
Financial correctness

AirbnbGlobal Payout Platform

Payouts to hosts across processors and banks

Worked on services responsible for sending payouts to hosts and tracking payout status across multiple payment processors and banking partners. Payouts had to handle failures from banks, delayed confirmations, and webhook inconsistencies.

We implemented payout retry workflows, status state machines, and better logging/tracing so support and operations teams could trace where a payout was stuck. We also improved payout success rates by refining retry timing and routing payouts through different processors depending on geography and failure patterns.

Technologies

Kafka
PostgreSQL
Java
Kotlin
AWS

Concepts

Payout workflows
Retries
State machines
Processor routing
Operations tooling

LyftRide Billing & Fare Calculation

Trip billing, pricing, and driver payouts

Worked on backend services responsible for calculating ride fares, fees, promotions, and driver payouts based on trip lifecycle events. Billing was triggered by trip completion events and required accurate pricing, taxes, and promotions.

We encountered issues with duplicate trip events causing duplicate charges, so we implemented idempotent billing keyed by trip ID. As trip volume increased, we also optimized slow MySQL queries and added indexes to keep billing queries performant during peak hours. We also built internal tools used by operations to reprocess trips and correct billing issues.

Technologies

Python
Java
MySQL
Redis

Concepts

Event-driven billing
Idempotency
Pricing
Receipts
Operational tooling

LyftEvent Processing Platform (Kafka)

Trip event pipeline for billing, notifications, and analytics

Helped migrate parts of Lyft’s trip processing workflow from synchronous service calls to an event-driven architecture using Kafka. Trip events such as trip_started, trip_ended, and payment_completed were published to Kafka and consumed by downstream systems including billing, notifications, and analytics.

We implemented retry handling, dead-letter queues, and monitoring for consumer lag and processing failures. This platform became the foundation for multiple downstream systems and allowed teams to build new features without tightly coupling to the trip service.

Technologies

Kafka
Python
Java
PostgreSQL
AWS

Concepts

Event-driven architecture
Consumers
Retries
DLQ
Schema evolution
Monitoring

Experience

Where I've built production systems

Payments, ledgers, payouts, and event-driven platforms—shipping under real reliability and scale constraints.

Airbnb

Senior Software Engineer, Payments Platform

Apr 2023 – Present

Subscription billing platform

Backend systems for recurring subscriptions, invoicing, and payment collection.
Designed a monthly billing workflow that generates invoices and charges customers; moved invoice generation from synchronous processing to an asynchronous Kafka workflow to reduce peak load on Postgres.
Implemented payment retry and dunning for failed payments with idempotency keys to prevent duplicate charges during retries.
Added database indexes and backfilled historical invoices so new billing queries could run without locking tables during peak hours.
Partnered with finance so invoice and payment data matched reporting requirements.

Java
Kotlin
PostgreSQL
Kafka
Redis
AWS

Financial ledger & reconciliation

Built a ledger service to track billing events, payments, refunds, and payouts.
After duplicate ledger entries during a retry storm from a downstream processor, shipped idempotent write paths and reconciliation jobs to detect and fix mismatches.
Designed append-only ledger tables and balance logic in PostgreSQL.
Built reconciliation jobs comparing ledger data with processor reports and discrepancy tooling for finance.
Wrote design docs and led architecture and data model reviews.

Java
PostgreSQL
Redis
Kafka

Global payout processing

Services for sending payouts to hosts and tracking payout status through the pipeline.
Payout retry workflows and failure handling when bank transfers or processors failed.
Status tracking so support and ops could trace stuck payouts.
Improved success rates by refining retry timing and multi-processor routing.
Incident investigation with stronger logging and tracing for payout workflows.

Java
Kotlin
PostgreSQL
Kafka
AWS

Lyft

Software Engineer, Backend / Distributed Systems

Jun 2018 – Apr 2023

Ride billing & fare calculation

Backend services for ride fares, fees, promotions, and driver payouts.
Processed trip completion events into billing records after trips finished.
Idempotent event processing when duplicate trip events risked duplicate charges.
Optimized MySQL queries and added indexes as trip volume grew and peak-hour latency appeared.
Internal tools for operations to reprocess trips and fix billing issues.

Python
Java
MySQL
Redis

Trip event processing platform

Migrated trip processing from synchronous services toward Kafka-backed async workflows.
Consumers for trip lifecycle events such as trip started, trip ended, and payment completed.
Retry handling and dead-letter queues for failed event processing.
Monitoring and alerts for processing delays and failures.
Foundation for downstream billing, notifications, and analytics.

Kafka
Python
Java
PostgreSQL
AWS

Skills

Tools and domains I ship with

Backend-first, with production experience across payments, ledgers, and high-throughput event pipelines.

Languages

Java
Kotlin
TypeScript
Python
SQL

Frontend

React
Next.js
TypeScript
HTML
CSS

Backend

Node.js
Spring Boot
FastAPI
REST APIs
Microservices

Databases

PostgreSQL
MySQL
Redis
Elasticsearch