AI Infrastructure Engineer

KAISHI PARTNERS PTE. LTD.; Singapore; On-site; SGD 120,000 - 180,000; Full time

Job summary

KAISHI PARTNERS PTE. LTD. in Singapore is expanding its engineering team, seeking an Infrastructure Engineer to own and evolve the backend infrastructure for AI agents and stateful workflows. You will be the first infra hire, collaborating with founders on architecture and reliability.

You will build Python-based infrastructure for agent execution, design recovery for interrupted workflows, and connect conversations to outcomes while focusing on latency, costs, and robust deployment practices.

Qualifications

• Strong software engineering skills in Python and backend systems.
• Experience owning production infrastructure and reliability trade-offs.
• Practical understanding of asynchronous processing, state management, APIs, and data persistence.
• Ability to balance rapid delivery with solid architecture.
• Comfort working independently in a small evolving team.

Responsibilities

• Build and evolve Python infrastructure for agent execution, persistent state, model routing, and tool use.
• Design recovery mechanisms for interrupted workflows, including retries and failure handling.
• Develop event and execution-history pipelines linking conversations, decisions, actions, and outcomes.
• Establish deployment, monitoring, and debugging practices for a small team.
• Improve latency, reliability, and model usage costs as the product develops.
• Make pragmatic architectural decisions and own systems through production failures.

Skills

Python

Distributed backend

Production infra

Async processing

State management

APIs

Data persistence

Independent worker

Job description

Our client is an early-stage, venture-backed AI startup building a personal agent that understands users’ preferences, helps them discover relevant products, and follows through on their decisions. Combining AI agents, personalisation, and commerce, the team aims to create a service that learns from real customer experiences and becomes more useful over time.

They are expanding their engineering team in Singapore, offering early hires substantial ownership of the product and technical architecture, working closely with the founders across AI infrastructure and full-stack development. You will be the first Infrastructure Engineers to join their team

The opportunity

You will build the infrastructure that makes AI agents dependable and affordable to run. These agents need to retain context across conversations, execute actions reliably, recover from interruptions, and connect their decisions to outcomes that may arrive weeks later.

This is a hands‑on software engineering role with substantial ownership of architecture and production reliability. You will work closely with product engineers and researchers to turn emerging agent capabilities into systems people can rely on.

What you’ll do

• Build and evolve the Python infrastructure supporting agent execution, persistent state, model routing, and tool use.
• Design recovery mechanisms for interrupted workflows, including retries, failure handling, and protection against duplicate actions.
• Develop event and execution-history pipelines that connect conversations, agent decisions, actions, and outcomes.
• Establish deployment, monitoring, and debugging practices that help a small team operate effectively.
• Improve latency, reliability, and model usage costs as the product develops.
• Make pragmatic architectural decisions and take ownership of systems through production failures and changing requirements.

What you’ll bring

• Strong software engineering skills in Python and experience building distributed backend systems.
• Experience owning production infrastructure, investigating incidents, and making reliability trade-offs.
• Practical understanding of asynchronous processing, state management, APIs, and data persistence.
• The ability to balance immediate delivery with sound technical foundations.
• Comfort working independently in a small team where requirements and architecture are still evolving.

Useful additional experience

• Agent runtimes, model API integrations, or ML infrastructure.
• Durable workflows, event-driven systems, or execution tracing.
• Measuring and optimising inference cost and latency.