About the Role
We are seeking an SDE-3 Backend Builder to own the CRM Opportunities product end-to-end. Opportunities is the deal and pipeline engine at the heart of our CRM — the system of record for every revenue motion our customers run, from lead conversion to closed-won. This role emphasises deep backend ownership: designing data models, APIs, and distributed workflows that power pipelines, stages, automations, reporting, and integrations at scale. You will operate with high autonomy, turning ambiguous problems into production systems, with enough frontend fluency to ship the occasional UI surface without waiting on anyone. Opportunities is also where a growing set of AI-assisted features live — things like surfacing what needs attention in a pipeline, summarising deal context, and suggesting next steps. To be clear about the shape of that work: we are not training or serving models. These features are built on top of existing AI capabilities, and the engineering problem is the product plumbing around them — assembling the right context from CRM data, keeping latency and cost sane, handling failure and non-deterministic output gracefully, and making the results trustworthy enough for a customer to act on. You do not need an ML background; you need the judgement to build a reliable product feature on top of a probabilistic dependency.
What you will do
Own the full lifecycle of the Opportunities backend: translate business and customer needs into scalable, production-grade services
Design and evolve the data model for pipelines, stages, deals, and associated entities — getting consistency, throughput, and query patterns right
Build and scale APIs, event-driven workflows, and background jobs that handle high-volume writes, automations, and integrations
Tackle distributed systems concerns: idempotency, eventual consistency, race conditions, hot tenants, and noisy-neighbor isolation
Improve performance and reliability of large pipelines (millions of opportunities per workspace) through indexing, sharding, caching, and query optimization
Build the AI-assisted surfaces of Opportunities on top of existing AI capabilities — context assembly from CRM data, prompt and response contracts, caching, cost and latency budgets, graceful degradation when a call is slow, wrong, or unavailable, and a feedback loop for measuring output quality
Ship end-to-end when needed — including UI changes in Vue — without blocking on cross-functional handoffs
Instrument the system: logs, metrics, traces, and SLOs that make production behavior legible
- Drive incident response, postmortems, and the engineering hygiene that keeps a high-traffic product healthy
- What you will need:
- 4+ years building and operating production backend systems at scale
Strong fundamentals in data modeling (SQL and NoSQL), API design, and distributed systems
Hands-on experience with Node.js or Go in production
Track record of owning services end-to-end — design, build, deploy, on-call
Comfort working independently in ambiguous problem spaces with high ownership
- Solid grasp of scalability, performance, and reliability trade-offs
- Preferred Qualifications:
- Experience building CRM, sales, pipeline, or workflow products
Worked on multi-tenant SaaS at scale (sharding, tenant isolation, fair-use limits)
Familiarity with event streaming (Kafka/Redis Streams), queues, and async processing
Exposure to MongoDB, Postgres, ElasticSearch, ClickHouse, or similar at non-trivial scale
Have shipped a customer-facing feature backed by an LLM or similar AI API — and have opinions on where it worked and where it did not. No modelling or ML-infrastructure background required
Frontend fluency in Vue.js — enough to ship a feature without a frontend partner
- Strong product instincts and a portfolio or GitHub of shipped work
- Tech Skill Set:
- Node.js (NestJS), Go, Vue.js, MongoDB, Redis, Kafka / Pub-Sub, ElasticSearch — running on a distributed microservices architecture over Istio service mesh, with AI-assisted features built on top of existing AI/LLM APIs.
- What Success Looks Like:
- Ships Opportunities features from idea to production with minimal cycle time
Builds services that stay reliable as workspaces grow from thousands to millions of deals
Makes sound trade-offs between speed, quality, and long-term maintainability
Ships AI-assisted features that customers actually trust — useful often enough to matter, and well-behaved when the underlying call is not
Raises the bar on data correctness, API design, and operational excellence for the product
Continuously sharpens the customer experience through iteration and instrumentation
