Senior SDE - Conversations

Job Summary

Were hiring backend engineers who treat that as a design constraint, not a fun fact.

This isnt a role where youll implement someone elses spec. Youll be handed problems that dont have a known answer yet - a datastore buckling under index growth, a queue that needs to absorb 10x burst traffic without dropping a message, a state machine with millions of concurrent executions. Youll figure it out. And youll own it.

Team & System Overview

The Conversations organization powers multi-channel communication across SMS, Email, WhatsApp, and DMs - handling over 2B messages each month. The Threads & Composer team owns the inbox and composer surfaces, and the backend services that deliver and store those interactions. You ll work across Node.js microservices, GCP-based infra (GKE, Pub/Sub, Cloud Tasks), and MongoDB/Firestore storage layers, ensuring every message and UI state stays consistent, fast, and recoverable.

Responsibilities

• Design and build Node.js / Go services on critical paths - high throughput, low latency, zero tolerance for data loss
• Own systems end to end: architecture, design docs, code, tests, rollout, and production health
• Model and optimize data layers across SQL and NoSQL stores handling high read/write volumes and atomic updates
• Build queueing, caching, and rate-limiting layers that keep the platform graceful under burst traffic
• Instrument everything - logs, metrics, traces - so you know your system is healthy before anyone asks
• Debug production incidents, ship the fix, drive the post-mortem to closure
• Review code, raise the bar, and pull the engineers around you up with you

Requirements

• 4+ years of backend engineering experience, with deep hands-on work in Node.js and/or Go
• Queueing experience - you've designed with Pub/Sub, Kafka, RabbitMQ, Cloud Tasks, or similar, and you understand delivery semantics, ordering, and backpressure
• Complex distributed systems experience - consistency trade-offs, idempotency, fault tolerance, systems that stay correct under partial failure
• Hands-on with Redis or other in-memory data stores - caching strategies, data structures, and their failure modes
• Strong with SQL or NoSQL databases - schema design, indexing, and query optimization at high volume
• You write clear design docs and rigorous test cases as a habit, not on request
• Champion of AI-assisted engineering - you make agents produce accurate, production-quality code, fast

Technical Environment

• Languages: Node.js (TypeScript), Go
• Storage: MongoDB, Firestore, Redis, ElasticSearch, ClickHouse
• Infra: GCP - GKE, Pub/Sub, Cloud Tasks
• CI/CD: GitHub Actions, Docker, Kubernetes

On AI

Were past the debate. AI is part of how we build - and we expect you to be better at it than most.

That means making agents do real work: producing code thats accurate, tested, and slop-free - fast. You know when to hand a problem to an agent and when to take the wheel. You review AI output with the same rigor as a junior engineers PR, and you ship at a pace that engineers without this skill cant match.

If youre still copy-pasting from a chat window and hoping, this isnt the seat for you.

What You're Built For

• Youve solved problems nobody handed you a playbook for - ideally in a 0 1 environment where the system didnt exist until you built it
• You understand systems in depth - not just the API surface, but what happens under load, at the tail, when the network partitions
• You write design docs people actually read - clear trade-offs, honest risks, a real recommendation
• You treat test cases as part of the design, not a chore after the fact