Blockit
Software Engineer, Infrastructure
San Francisco
Sponsorship not specifiedDetected 181 days ago
PostgreSQLRedisVector DatabasesDevOpsKafkaLLMsAgentic AILogistics
About the role
- Think of this as a product engineer role - but the product is everything underneath.
- For a stateful, multiplayer AI agent that takes real actions in the world, infrastructure is the user experience.
- When a meeting gets scheduled in 800ms instead of 4 seconds, when an email never gets dropped, when our agent recovers gracefully from a flaky third-party API - that's a feature users feel.
Responsibilities
- Own, operate, and evolve our core data infrastructure: PostgreSQL, ClickHouse, and async processing pipelines - measuring success in user-facing terms (p95 scheduling latency, agent action success rate, message delivery reliability)
- Own observability end-to-end: metrics, logs, traces, alerting, dashboards, and the on-call / pager rotation - including what we alert on and what we don't
- Design, manage, and optimize our LLM infrastructure (gateway, evaluation pipelines, agent observability) for reliability, performance, and cost
- Lead infrastructure migration decisions and execute them: e.g. evaluating pg-boss vs. managed queue alternatives, weighing whether to move off Postmark, introducing Redis / Kafka / similar when the time is right
- Partner directly with product engineers to ship features - the line between "infra work" and "product work" should be invisible here
- Own DevX: shape how engineers leverage AI in their day-to-day workflow - local environments, Claude Code conventions, agent tooling, eval harnesses, anything that compounds team velocity
- Own DevOps: deploy pipelines, environment management, capacity planning, infra cost
Requirements
- 4+ years of software engineering experience, with significant time on backend or infrastructure
- Product sensibility - you reach for user-facing metrics first, and you can tell the difference between an infra problem that matters and one that doesn't
- Deep with databases - you can read query plans, find the bottleneck, and know when to fix the query vs. fix the schema vs. fix the architecture
Skills
- deploy pipelines, environment management, capacity planning, infra cost
Company info
- The goal is that we catch problems before customers do, and that when something does break, the engineer paged knows exactly where to look.
Apply directly at Blockit →Create a free account for alerts like thisView Blockit immigration profile
This listing is sourced directly from Blockit's careers page and normalized into a canonical job model.