全职 Machine Learning Engineer 工作, 薪水, Confidential Jobs Federal Territory 公司招聘中 - Ricebowl

Position Overview

We are seeking a Machine Learning Engineer to own the productionisation of computer vision models — taking trained models from research to reliable, optimised systems serving in production across both cloud infrastructure and on-device edge hardware. The role sits between model science and software engineering, requiring both ML systems depth and strong engineering discipline.

Key Responsibilities

Model Optimisation & Deployment

• Deploy and maintain model serving infrastructure across both cloud (AWS EC2/SageMaker) and edge (Jetson) environments

• Optimise models for their deployment target: TensorRT engine export, INT8/FP16 quantisation, and ensemble fusion for edge; scalable serving infrastructure for cloud

• Validate accuracy-speed trade-offs across environments; define and enforce latency/throughput acceptance criteria before production rollout

• Manage model versioning and coordinated rollout across a mixed fleet of cloud-served and edge-deployed devices

Inference Pipeline Engineering

• Build and maintain end-to-end inference pipelines from model input to structured output

• Ensure inference reliability, error handling, and graceful degradation in field conditions

• Define integration contracts (input/output specs, performance envelopes) for downstream software systems

Evaluation & Active Learning

• Design and maintain evaluation frameworks: per-class performance, regression testing, benchmark reproducibility

• Build and maintain active learning pipelines to surface high-value samples for annotation

• Define annotation standards and review annotation quality in collaboration with the CV team

Collaboration

• Work closely with CV Scientists on model handoff requirements; with Software Engineers on API integration

• Communicate inference system performance and limitations to technical and non-technical stakeholders

Required Qualifications

Education & Experience

• Bachelor's or Master's in Computer Science, Electrical Engineering, Machine Learning, or related field

• 4+ years of experience in machine learning engineering, with production deployments to your name

Technical Skills

• Python — production-quality code; not just experimentation scripts

• Strong working knowledge of at least one major deep learning framework (PyTorch preferred)

• Solid understanding of computer vision fundamentals: object detection, image classification, model evaluation metrics (mAP, precision/recall, IoU)

• Experience with model serving and inference systems — loading, scripting, and serving trained models via REST APIs or equivalent

• Familiarity with MLOps practices: model versioning, experiment tracking (MLflow or equivalent), reproducible benchmarking

• Experience working with image datasets: understanding of annotation formats, dataset splits, class imbalance, and data quality issues

• AWS (S3, EC2, SageMaker); Docker; Git — comfortable across the full development-to-deployment workflow

• Able to write clear integration documentation: API contracts, performance envelopes, known failure modes

Soft Skills

• Methodical — validates before shipping; distinguishes a noise result from a real improvement

• Communicates ML system performance and limitations clearly to non-ML stakeholders

• Takes ownership through to production — does not consider work done at model handoff

Good to Have

• Model compression for edge: TensorRT, ONNX export, INT8/FP16 quantisation

• Model serving infrastructure: Triton Model Server or equivalent

• Deployment to edge hardware: Jetson Orin or comparable resource-constrained device

• Active learning pipeline design and annotation tooling (Label Studio or equivalent)

• Familiarity with multi-label or ensemble model architectures

• Experience with object tracking algorithms (ByteTrack, Kalman filter-based approaches)