Software Engineer, Data Infrastructure (Staff)

ABOUT LIGHTFIELD Lightfield is an AI-native CRM that assembles itself from your email, calendar, and meetings. It captures every interaction and turns it into organized context: accounts, tasks, follow-ups, and insights, so nothing slips through the cracks. We’re rethinking CRM from first principles. Instead of forcing teams to maintain rigid systems, Lightfield learns from how companies actually work, adapting, automating, and surfacing the insight that drives growth. We’re building the CRM platform we always wished existed: fast, intelligent, and genuinely helpful. The market agrees. Over 5,000 companies have used Lightfield since we launched in November, making us the fastest-adopted product in our category. We are backed by Greylock, Lightspeed, and Coatue, and our founders previously built Tome, a generative AI presentation product used by over 25 million people. Before Lightfield, our team worked on Llama, Instagram, Facebook Messenger, Pinterest, Google, and Salesforce. ABOUT THE ROLE We're building the infrastructure foundation for a fast-growing AI product company serving thousands of customers: - Our Postgres fleet serves 5B+ queries a month — roughly 2,500 QPS steady state, with sustained spikes past 25,000 QPS — and database workload more than doubled last month. - Redis sustains ~50,000 commands per second behind a job platform that executes 10M+ background job runs a day across ~170 queues. - We ingest tens of millions of emails and calendar events a month. That growth creates scaling pressure across backend systems, infrastructure, and data infrastructure. We're hiring a staff-level engineer who spikes in data infrastructure but is excited to work across backend systems, infrastructure, and product-facing data problems. The work is close to the product, close to customers, and close to production. As a Software Engineer, Data Infrastructure, you'll build the next generation of our data systems. We got remarkably far on a deliberately simple stack: Postgres as the system of record, a sharded transactional outbox for change events, Redis-buffered sync into Typesense for search, BullMQ for processing, and Postgres-backed customer-facing analytics with per-organization row-level security. The next phase is evolving that pragmatic foundation into best-practice data architecture: change data capture, event modeling, schema design, query performance, freshness guarantees, and the right boundary between transactional and analytical workloads. The system of record itself is unusual. Customers define their own objects, attributes, and relationships at runtime, so the core data model is a schema-flexible, graph-shaped store: entity-attribute-value with typed edges, versioned attribute values, and relationship history. That makes schema design, indexing, and query performance genuinely hard problems rather than routine tuning. The surface area is wider than analytics: customer-facing dashboards, historical and audit data, datasets that power pipeline-generation products, and evaluation data that measures our AI agents. This is data infrastructure work, not a BI or dashboarding role. It's a good fit for someone who likes high-volume data systems, pragmatic architecture decisions, and building foundations that product and engineering teams can actually depend on. This role can be based in San Francisco or Cambridge. In San Francisco, you'd work from our HQ alongside the founders and most of the engineering team. In Cambridge, you'd join an initial group of staff-level engineers at our new, infrastructure-focused Kendall Square site, working alongside one of our most senior infrastructure engineers. We aim to build the site and organization around this group as the company scales. WHAT YOU'LL DO - Scale the analytics engine behind customer-facing dashboards, tackling query performance under row-level security, workload isolation, read architecture, and observability as data volume grows. - Design the ingestion paths, event mod