Principal Machine Learning Engineer
Bjak · Remote · Singapore ·
COMPENSATION
€•• – ••k
Listed on the employer's job board (Ashby)
- Principal level role
- Fully remote
- Python + PyTorch/JAX
- GPU-based systems
- High ownership
Join a small, high-talent-density team building proactive AI applications for email, notes, and tasks. The goal is to bring intelligence to daily workflows with minimal prompting, focusing on reliability and real-world task completion.
As a Principal Machine Learning Engineer, you will own the execution layer of the company's intelligence. You will bridge the gap between research and product by creating scalable systems for training, evaluation, and inference.
What you'll do
- Own end-to-end ML systems from data and training to deployment
- Build and evolve training and fine-tuning pipelines for large models
- Design evaluation systems for capability, robustness, and safety
- Architect high-performance inference systems optimizing latency and cost
- Create data pipelines for high-quality real-world and synthetic data
- Establish infrastructure for deploying, monitoring, and improving models
- Partner with research and engineering to turn capabilities into products
- Make pragmatic technical trade-offs based on real-world performance
What you'll need
- Experience shipping production ML systems, not just research prototypes
- Strong understanding of large-model training, fine-tuning, and inference
- Solid software engineering and systems fundamentals
- Experience operating ML workloads at scale on GPU-based systems
- Ability to navigate ambiguous problems with strong technical judgment
- Bias toward experimentation, measurement, and shipping code
- High standards for correctness, reliability, and production quality
Stack
- Python
- PyTorch
- JAX
- GPU-based training and inference systems
How you'll work
- Small, hands-on team with broad individual ownership
- Fast decision-making balancing speed with engineering fundamentals
- Interviews conducted via virtual meetings or onsite
- Prompt hiring decisions after 3 to 4 interviews