Job Description

Kaishi is currently with a fast growing AI-focused start-up building enterprise software for regulated industries, including banking, financial services and insurance.

What they do

The firm build AI-powered tools that support customer-facing teams through training, real-time assistance and automated workflows. Their enterprise clients operate across markets and languages. Key engineering challenges include latency, reliability, testing non-deterministic systems, meeting stringent security requirements and delivering natural multilingual AI interactions.

Why this role exists

The business has expanded from an initial product to multiple products serving enterprise clients. Much of this growth has relied on client-specific implementations. Shared capabilities such as tenant configuration, identity, model access and evaluation now exist in several versions, increasing the engineering effort required for each new client.

We want a common platform: shared services, configuration per client instead of code per client, and eventually a builder that lets us assemble new use cases. You would own the platform core, extract shared capabilities while products continue shipping, and set the technical bar for a distributed engineering team. This is a hands-on individual contributor role with weekly coding responsibilities, rather than a management or advisory-only architecture position.

What you would do

. Build shared capabilities for configuration, identity and permissions, tenant isolation, model and retrieval access, and evaluation.

. Extract these capabilities from existing products without stopping delivery, and drive adoption across multiple teams.

. Make architecture and migration decisions across repositories, from design through production operation.

. Set and uphold standards for ownership, review, testing, security and release, while helping teams deliver under pressure.

. Provide safe ways to vary behaviour per client without client-specific forks or hard-coded conditions.

. Work with infrastructure engineers on deployment, CI/CD and observability.

. Ensure every important system has a second owner, including the systems you build.

First three months

Month one: Get to know the codebase and platform work already under way agree the migration sequence with technical leads become second owner of one critical system.

Month two: Extract the first shared capability and run it in production for one product, with tests, an owner and an adoption plan for the next product.

Month three: Have a second product adopting the capability or migrating document and agree the next two quarters of platform work establish a named second owner for what you have built.

What we need from you

. 8 or more years building production backend systems, with Staff-level experience: architecture across several teams, end-to-end migrations, and adoption driven in teams that did not report to you.

. Deep production experience with TypeScript and Node.js. Comfortable working in Python and reviewing code outside your main language.

. Practical multi-tenant SaaS experience: designing, testing and operating controls that prevent one client's data from reaching another.

. Experience decomposing a monolith or extracting shared services while keeping it in production, with clear examples of sequencing, trade-offs and lessons learned.

. Infrastructure judgment covering Kubernetes, GCP or AWS, CI/CD and observability. You can design for and review these systems, even if you do not operate them directly.

. Strong security and data-handling instincts, ideally gained in a regulated environment.

. Clear written communication for a distributed team working across time zones.

. Ability to push back on unsafe shortcuts under pressure while maintaining trust.

Nice to have

. Experience in banking, insurance or another regulated industry, including security reviews, SOC 2 or Singapore's OSPAR.

. LLM application experience: model gateways, retrieval, prompt versioning, evaluation harnesses and testing non-deterministic output.

. Experience with internal developer platforms or self-service configuration systems.

. Practical use of AI coding tools and sound judgment about their benefits, limitations and review requirements.