Senior Backend Engineer

About the role Peec AI is building the data plane for AI visibility: millions of chats, scrapes, citations, sources, model responses, and customer-facing insights running through a fast-moving backend. This role is for an engineer who wants to own the systems behind that. You will not just add endpoints or close tickets. You will make our scraping, chat-processing, source-ingestion, and analytics infrastructure more reliable, cheaper, faster, and easier for the rest of the team to reason about. The best fit looks like someone who can debug a provider outage in the morning, design a better queueing model before lunch, ship a TypeScript service change in the afternoon, and write the postmortem or follow-up ticket clearly enough that the next engineer immediately knows what to do. We are especially interested in people who have built or operated systems with real constraints: rate limits, flaky third-party APIs, queues, webhooks, retries, distributed locks, data migrations, observability gaps, customer pressure, and cost ceilings. What we’re looking for 5+ years of backend, platform, infrastructure, or distributed systems engineering experience in a production environment. Deep hands-on TypeScript experience. We run TypeScript services on Bun, and we care about people who can design clean, reliable backend systems rather than only write application code. Strong understanding of queues, workers, retries, rate limits, idempotency, distributed locks, webhooks, and failure handling. Experience operating systems under real production pressure: incidents, provider outages, customer escalations, noisy alerts, slow dependencies, and incomplete data. Comfort with cloud-native infrastructure. Our stack includes GCP, Kubernetes, Cloud Tasks, Pub/Sub, RabbitMQ, Redis, Postgres, Firestore, ClickHouse, Prometheus, Grafana, and OpenTelemetry. Ability to debug across layers: application code, logs, metrics, queues, provider APIs, databases, network behaviour, and deployment state. Product-minded judgement. You can separate internal detail from customer-safe explanation, and you understand when a “small infra change” has customer, security, or cost implications. High ownership and low ego. We want people who take responsibility for ambiguous systems, ask sharp questions, and make the team better through clear thinking and kindness. A bias toward durable fixes. You do not stop at “restart the pod” if the real issue is retry amplification, missing backpressure, an unsafe parser assumption, or a provider limit we do not model. Tech stack Languages: TypeScript, Python Runtime/frameworks: Bun, Elysia, Node-compatible services Cloud/infrastructure: GCP, Kubernetes, Cloud Tasks, Pub/Sub, Firebase/Firestore, Postgres, ClickHouse, Redis, RabbitMQ Observability: Prometheus, Grafana, OpenTelemetry, structured logging Product/data domain: LLM chats, AI search, source scraping, citation extraction, provider integrations, customer-facing analytics AI/model providers: OpenAI, Anth