全职 Principal Machine Learning Engineer 工作, 薪水, Grab 公司招聘中 - Ricebowl

Get to Know the Team

The AI Platform (AIP) team builds and operates the core ML and AI infrastructure that powers Grab. Our stack spans model serving, ML pipelines, data serving, AI infrastructure, and Applied Research. Together, AI Platform serves hundreds of data scientists and ML engineers across Grab, and is the foundation for everything from fraud detection and search ranking to foundation model efforts, adaptive experimentation, LLM fine-tuning, and the next generation of agent-driven products.

Get to Know the Role

As a Principal Machine Learning Engineer on AI Platform reporting to the AIP Head of Engineering, you'll be the senior technical anchor and solution architect for the platform.

Your mandate is twofold: (1) Raise the ceiling and raise the floor. You'll partner with the Head of Engineering, AI Platform and Applied Research team to shape the forward roadmap, evaluate SOTA techniques worth productionizing, and drive integrations that connect AIP's capabilities into a coherent end-to-end experience. (2) You'll ensure the teams who depend on AIP can actually succeed on it — translating user pain into platform requirements, unblocking complex adoption cases, and getting hands-on where it matters. You'll take on workstreams spanning cross-entity platform consolidation, large-scale training throughput and reliability, and faster, higher-quality model iteration for AIP's most important users. This is an individual contributor role for someone who thrives at the intersection of platform engineering, applied ML, and user empathy — and who is equally comfortable writing a design doc, debugging a training job, and pairing with other teams to land their next model on the platform.

The Critical Tasks You Will Perform

• Solution Architecture for AIP Users: Partner directly with Data Scientists and ML engineers across the company to design end-to-end solutions on AIP. Be the senior technical escalation point for complex adoption cases.
• Large-Scale Training: Drive the state of large-scale training on AIP — throughput, reliability, cost, and developer experience. Advise and contribute hands-on across foundation model training, RL, simulation, and LLM fine-tuning workloads.
• Faster Iteration and Model Quality: Attack the end-to-end loop from idea to shipped model — data, training, evaluation, deployment, monitoring. Drive measurable reductions in iteration time and measurable gains in model quality for AIP's highest-value use cases.
• Platform Integration: Design and drive integrations across AIP surfaces (model serving, ML pipelines, data serving, AI infra, AI Automation tooling) so users experience a coherent platform rather than a collection of services.
• User Experience Translation: Convert pain points surfaced through embeddings, support channels, and direct user work into functional requirements for AIP teams; propose cross-platform solutions that raise the bar for the DS and MLE personas.
• Enablement at Scale: Produce reference architectures, patterns, and opinionated best-practice guidance so the next hundred ML use cases land on AIP cleanly, without requiring bespoke platform-team involvement every time.
• Strategic Roadmap Definition & Mentorship: Stay current with SOTA across ML infrastructure, LLM serving, training systems, and RL; partner with the AIP HoE to decide what Grab should adopt, build, or skip. Raise the technical bar across AIP through code, design reviews, written artifacts, and direct mentorship of senior engineers across the org.

What Essential Skills You Will Need

• Advanced MLOps & ML Platform Engineering: Expert-level mastery of ML lifecycle platforms (e.g., Kubeflow, MLflow, Triton, TorchServe) and distributed training frameworks (PyTorch, Ray, Horovod).
• Distributed Systems & Infrastructure: At least 8 years of experience in Kubernetes, containerization, and high-performance computing clusters (GPUs/TPUs). Experience optimising large-scale data and model pipelines.
• Architecture Design: Outstanding system design capability for available, scalable, and secure multi-tenant platform services.
• AI/LLM System Experience: Hands-on experience with LLM orchestration, fine-tuning infrastructure, or serving optimization (vLLM, TensorRT-LLM).
• Innovation & AI Fluency: A learning mindset to evaluate and implement state-of-the-art (SOTA) infrastructure paradigms, guiding the team on what to build versus what to skip.
• Adaptive Execution & Ownership: You can operate independently in high-ambiguity environments, taking full end-to-end accountability for complex system integrations and platform consolidation.
• Coaching with Care: Commitment to raising the engineering bar. Mentor senior engineers and foster a culture of technical excellence and collaboration.