Software Engineer — AI-Native Agentic Systems

United States Digital Space LLC; Singapore; On-site; SGD 120,000 - 180,000; Full time

Job summary

United States Digital Space LLC in Singapore seeks software engineers to join the Acceleration Platform team. You will design, develop, test, deploy, and maintain AI-native agentic systems that eliminate developer toil at enterprise scale.

You will blend prompt engineering, RAG, and optimization to build highly capable AI applications. As a key member, you will help set the bar for AI engineering excellence, mentor peers, and ensure distributed systems are secure-by-default and highly

Qualifications

• Bachelor’s degree or equivalent practical experience.
• 2 years of experience with software programming in Python or C++.
• 1 year of experience in a core ML domain, including NLP, CV, speech, RL, or related areas.
• 1 year of experience with ML infrastructure (training, inference, deployment, evaluation, data processing).

Responsibilities

• Design and build AI-native agentic systems that automate complex workflows.
• Integrate prompt engineering, RAG, and optimization for capable AI applications.
• Create evaluation pipelines to measure AI performance and ensure secure-by-default behavior.
• Troubleshoot non-deterministic AI behaviors and refine context for reliable systems.
• Accelerate productivity using AI coding assistants and collaborate with global teams.

Skills

Python

C++

ML domain experience

ML infrastructure

Education

Bachelor’s degree or equivalent practical experience

Job description

United States Digital Space LLC in Singapore seeks software engineers to join the Acceleration Platform team. You will design, develop, test, deploy, and maintain AI-native agentic systems that eliminate developer toil at enterprise scale.

You will blend prompt engineering, RAG, and optimization to build highly capable AI applications. As a key member, you will help set the bar for AI engineering excellence, mentor peers, and ensure distributed systems are secure-by-default and highly