Senior/Staff Software Engineer, Data Platform

About the role

About Evervault

Evervault offers modular payments infrastructure for companies building their own payments stack. We keep raw card data off your systems, cut PCI compliance work, improve acceptance, mitigate fraud and chargebacks and decrease your cost of payments, so you keep full control. Trusted by Ramp, Rippling, Modern Treasury and Overwolf; Series B, led by Ribbit Capital.

Role Overview

You’ll be the first engineer dedicated to our data platform, and you’ll design and build it largely from scratch. You’ll own the platform that collects our event data, models it, and serves it to customers with roughly 15-second freshness, including building our time-series foundations. This is a senior, hands-on role: you’ll set the architecture, write most of the code and set the bar for the data team that grows around you. You’ll work closely with our VP of Engineering, Product and our leadership team. Product decides which data products to build; you own how they’re built and what the data can support.

What You’ll Be Doing

- Designing and building real-time ingestion for three sources: events our platform already emits, new instrumentation in our product, and data our customers send us

- Rebuilding our event streaming layer, with the authority to choose how

- Modelling structured and time-series data in ClickHouse and Databricks so aggregations, trends and benchmarks are fast and cheap at scale (our current defaults, though you can make the case for change)

- Designing the external data contracts and APIs customers build on, including versioning, reliability and isolation of each customer’s data

- Defining the event contracts our product engineers instrument against

- Owning the platform in production: monitoring, data quality, cost and on-call for the systems you build

Your First Year

In your first few months you’ll assess what we have, decide the platform architecture, and scope our first data product with Product and Engineering. After that, you’ll rebuild the streaming layer and ship the first external product on the new platform, most likely aggregated analytics or direct data access for customers. Later in the year you’ll help us hire a domain expert who decides which signals and benchmarks are worth building. Machine learning is out of scope for now; it comes once we have enough clean historical data.

Requirements

- Production experience building real-time or near-real-time ingestion with event streaming, such as Kafka, Kinesis or Redpanda

- Deep experience with ClickHouse or a similar analytical database, including modelling time-series data for fast aggregation at volume

- A track record of shipping an external, multi-tenant data product or data API that customers depended on (internal reporting pipelines alone won’t cover this)

- Experience migrating a messy live system to a new one without downtime or data loss