GenAI Systems Architect for Production LLMs

ONEBYZERO PTE. LTD.; Singapore; On-site; SGD 120,000 - 190,000; Full time

Job summary

ONEBYZERO PTE. LTD. seeks a Deep Learning Architect with 4+ years of experience to design production-grade GenAI systems in Singapore. You will shape end-to-end LLM architectures, RAG pipelines, and multi-agent setups with a strong emphasis on reliability and performance.

You will review designs, optimize models for latency and cost, and collaborate with engineering, data, product, and client teams to deliver high-impact GenAI solutions on tight timelines, with hands-on coding required.

Qualifications

• 3–6 years in ML engineering or model development roles.
• Hands-on CPT/SFT/LoRA/QLoRA on LLMs.
• Strong Python, production-ready code.
• Experience with model evaluation pipelines.
• Cloud deployment with cost and latency awareness.
• Strong problem-solving and independent work.
• Review designs and communicate trade-offs clearly.

Responsibilities

• Design end-to-end LLM system architecture for production use cases.
• Pre-train, fine-tune LLMs and domain-specific models.
• Design and run evaluation pipelines to benchmark performance and cost.
• Optimize models for latency, throughput, and token efficiency in production.
• Integrate fine-tuned models into multi-agent pipelines.
• Define enterprise integration patterns for GenAI systems.
• Collaborate with cross-functional teams to deliver high-impact solutions.
• Contribute hands-on code and raise engineering standards.

Skills

ML engineering

LLMs

Python

CI/CD

Tools

Bedrock/OpenSearch

Weaviate/Milvus

Docker/Kubernetes

Job description

ONEBYZERO PTE. LTD. seeks a Deep Learning Architect with 4+ years of experience to design production-grade GenAI systems in Singapore. You will shape end-to-end LLM architectures, RAG pipelines, and multi-agent setups with a strong emphasis on reliability and performance.

You will review designs, optimize models for latency and cost, and collaborate with engineering, data, product, and client teams to deliver high-impact GenAI solutions on tight timelines, with hands-on coding required.