Principal AI Engineer

Autohand AI Ltd.; Singapore; On-site; SGD 180,000 - 240,000; Full time

Job summary

Autohand AI Ltd. seeks a Staff AI Engineer to drive design and training of cutting-edge language models for code understanding and generation. You will own data pipelines, evaluation benchmarks, and model integration with Fantail, Moa, and related tools.

You will mentor engineers, shape technical direction, and contribute to open-source projects where appropriate, impacting developer productivity and product reliability.

Qualifications

• 5+ years of software engineering experience
• Excellent written communication for async collaboration across time zones
• Citizen or authorised work permit in Australia, New Zealand, Singapore, Brazil, or the United States

Responsibilities

• Design and train small and large language models for code understanding and generation
• Build and maintain data pipelines for supervised fine-tuning and reinforcement learning from human feedback
• Develop evaluation benchmarks that reflect real-world developer workflows
• Collaborate with product engineers to integrate models into Fantail, Moa, Commander, and Evolve
• Write and review code in Rust, Python, and TypeScript for model serving and inference optimizations
• Participate in design reviews and contribute to technical roadmaps
• Mentor other engineers on AI/ML best practices and research reproducibility

Skills

Python

Tools

PyTorch

JAX

Rust

TypeScript

Job description

About the role

You will work directly on the frontier of autonomous programming. Our team builds Fantail (a small, fast coding model) and Moa (a large reasoning model) that power our developer tools. As a Staff AI Engineer, you'll design training pipelines, experiment with model architectures, and ship improvements that make our agents smarter and more reliable. This is not a research-lab role you'll see your work deployed to real users within weeks.

The AI research team is small and flat. You'll own significant parts of the model development stack, from data curation to evaluation. Your decisions directly affect product performance and developer productivity. We expect you to set technical direction, mentor junior engineers, and contribute to open-source when appropriate. The impact is high: you'll shape how the next generation of developers interact with AI.

What you'll do

• Design and train small and large language models for code understanding and generation
• Build and maintain data pipelines for supervised fine-tuning and reinforcement learning from human feedback
• Develop evaluation benchmarks that reflect real-world developer workflows
• Collaborate with product engineers to integrate models into Fantail, Moa, Commander, and Evolve
• Write and review code in Rust, Python, and TypeScript for model serving and inference optimizations
• Participate in design reviews and contribute to technical roadmaps
• Mentor other engineers on AI/ML best practices and research reproducibility

What we're looking for

• 5+ years of software engineering experience, with at least 2 years focused on large language models or generative AI
• Strong Python skills and familiarity with PyTorch or JAX
• Experience training or fine-tuning transformer models at scale
• Proven ability to write clean, maintainable code and review others' work
• Comfortable with ambiguity we often define problems as we go
• Excellent written communication for async collaboration across time zones
• You must be a citizen of the following country: New Zealand, Australia, Singapore, Brazil or US

Nice to have

• Experience with functional programming
• Published research at venues like NeurIPS, ICML, ICLR, or ACL
• Contributions to open-source ML frameworks or tools
• Background in software engineering tooling or developer experience
• Rust, Go or other systems programming languages

Preferred Candidates

You must be a citizen or have authorised work permit in one of the locations where we hire: Australia, New Zealand, Brazil, Singapore, or the United States.