Senior AI Engineer at Workato in Singapore

Job description

• Ultimately, Workato believes in fostering a flexible, trust-oriented culture that empowers everyone to take full ownership of their roles. We are driven by innovation and looking for team players who want to actively build our company.
• But, we also believe in balancing productivity with self-care. That’s why we offer all of our employees a vibrant and dynamic work environment along with a multitude of benefits they can enjoy inside and outside of their work lives.
• If this sounds right up your alley, please submit an application. We look forward to getting to know you!
• Business Insider named us an “enterprise startup to bet your career on”
• Forbes’ Cloud 100 recognized us as one of the top 100 private cloud companies in the world
• Deloitte Tech Fast 500 ranked us as the 17th fastest growing tech company in the Bay Area, and 96th in North America
• Quartz ranked us the #1 best company for remote workers

Responsibilities

• We're building the infrastructure layer that connects enterprise systems to AI: an MCP Gateway, an AI Gateway, and the services around them. These are the systems that sit between LLM providers and everything else — routing, auth, rate limiting, observability, protocol translation. We're looking for a senior engineer who understands both the systems layer and the AI protocol layer, and who can build production-grade services in Go and/or Ruby.
• Design and develop the MCP Gateway and AI Gateway — production services that mediate between applications, AI agents, and LLM providers. This means protocol-level work: MCP server/client implementations, request routing, streaming, tool-call proxying, authn/authz, and tenant isolation. You'll build the core infrastructure, not just applications on top of it.
• Build high-throughput, low-latency network services. You'll work close to the wire: TCP, TLS, HTTP/1.1 and HTTP/2, JSON streaming, connection pooling, backpressure. When latency matters, you'll know exactly where it goes.
• Own the data layer from the application side. Deep PostgreSQL knowledge — schema design, indexing strategies, query planning, transactions and isolation levels, connection management. You're not a DBA, but you can read EXPLAIN ANALYZE output and fix the query, not just add an index and hope.
• Design for concurrency. Worker pools, queues, graceful shutdown, backpressure, race-free shared state. You can profile a service under load (pprof, flamegraphs, query stats), find the bottleneck, and fix it.
• Drive observability for AI systems. Metrics, tracing, and logging that actually tell you what's happening — token usage, latency per provider, cache hit rates, failure modes, cost per request.

Requirements

• Senior-level experience (5+ years) in Go, Ruby, or both. Any combination works: deep Go, deep Ruby, or strong in both. What matters is that you've shipped and operated production services in at least one of them.
• Go candidates: you know the stdlib deeply and prefer it over frameworks. net/http, crypto/tls, context, goroutines and channels, the memory model. You've done performance optimization on real services — allocations, GC pressure, lock contention — and you understand networking (TCP, TLS, HTTP, JSON) in depth, not just through a framework's abstraction.
• Ruby candidates: strong Rails in production — you know where Rails ends and Ruby begins, you've tuned ActiveRecord rather than fought it, and you've built services that stay fast under load.
• PostgreSQL depth from an application developer's perspective. Query optimization, indexing, transactions, connection pooling, migrations at scale. You don't need to administer the cluster; you need to write code that treats it well.
• Concurrency and profiling as a practiced skill, not a bullet point. You've debugged a production incident with a profiler open.
• Familiarity with Kubernetes and containers. You can deploy, debug, and reason about your services in a containerized environment — resource limits, health checks, rolling deploys, networking basics.
• You've built production services with proper observability, deployment pipelines, and security. You know how to run reliable systems and debug distributed systems when things break.
• You've worked with LLMs at the protocol level — message structures, tool calling, streaming responses, caching strategies. You know what's happening on the wire, not just what the SDK abstracts away.
• Familiarity with MCP (Model Context Protocol) is a strong plus — ideally you've built or integrated MCP servers/clients and understand the transport and capability negotiation layers.
• You can integrate with OpenAI-compatible and Anthropic APIs directly, evaluate responses, understand token usage, and optimize for latency and cost.
• You can judge, audit, and verify LLM output. You catch subtle bugs, security issues, and hallucinated APIs before they ship — and you're fluent enough in the underlying systems to know why the output is wrong, not just that it's wrong.
• You learn fast and stay current. AI infrastructure is moving weekly; you follow it and can evaluate new protocols and approaches quickly.
• You participate in technical design discussions and code reviews, and you can explain complex concepts clearly to engineers and non-technical stakeholders alike.
• You're comfortable owning the full lifecycle: design, implementation, deployment, monitoring, and continuous improvement.

Skills

Required

• Go
• Ruby
• Rails
• PostgreSQL
• Kubernetes
• Containers
• Networking (TCP, TLS, HTTP, JSON)
• Concurrency
• Profiling
• Observability
• LLM protocol-level experience
• OpenAI-compatible APIs