Title: Principal AI Engineer - ML CV, AI.DA STC
Aero - 600 West Camp Road, SG
About the Role
We’re looking for a hands-on ML CV Engineer to lead the development and deployment of robust, production-grade computer vision pipelines. In this role, you’ll own the full lifecycle of CV models - from data curation and preprocessing, through model training and evaluation, to deployment, monitoring, and automated retraining.
You’ll play a critical role in ensuring our vision systems remain accurate, responsive, and scalable under real-world conditions. Your work will directly impact applications involving image classification, object detection, segmentation, and other visual inference tasks.
This is a role for someone who thrives in full-stack ML development, combining deep modeling expertise with disciplined engineering and deployment practices.
Key Responsibilities
End-to-End Vision Systems
- Build computer vision pipelines covering data ingestion, cleaning, augmentation, and preprocessing.
- Train and optimize CV models (classification, detection, segmentation) with PyTorch, TorchVision, and modern frameworks (YOLO, Detectron2, MMDetection, DINO).
- Automate evaluation workflows to benchmark performance and detect drift over time.
Production Deployment & Integration
- Deploy models with containerized environments (Docker, TorchServe, ONNX Runtime, BentoML) and expose via APIs (REST/gRPC).
- Collaborate with engineers to integrate models into larger platforms with reliability at scale.
Automation & Orchestration
- Design automated pipelines for data validation, retraining, and deployment (RPA).
- Implement workflow orchestration with Airflow, Prefect, or Dagster for scheduled training, monitoring, and failure recovery.
Monitoring & Reliability
- Monitor production performance, detect drift, and handle recovery gracefully.
- Build alerting and observability with Prometheus, Grafana, or OpenTelemetry.
Collaboration & Tooling
- Contribute to MLOps tooling for reproducibility, experiment tracking, and data versioning (MLflow, wandb).
- Work with AI Engineers to ensure clean integration with orchestration frameworks.
Must-Have Skills
- 6+ years of ML or CV engineering, including 3+ years building production-grade vision systems.
- Strong knowledge of CV tasks and architectures (classification, detection, segmentation).
- Proficient in PyTorch, TorchVision, Albumentations, and modern CV frameworks.
- Proven experience training and tuning models on real-world datasets.
- Skilled in production deployment (Docker, TorchServe, ONNX Runtime, BentoML, Kubernetes).
- Strong software engineering foundation: clean Python, Git workflows, testable architecture.
- Experience with ML orchestration tools (Airflow, Prefect, Dagster).
- Familiarity with monitoring and alerting systems for ML models.
What We Offer
- Small, agile team (5–6 engineers + interns) with autonomy and real ownership.
- Startup feel with a big company resources:
- International environment where the majority of the team and leadership is from startups or big international corporations (Lazada, Gojek, IBM) and from various countries.
- Low-bureaucracy, high-impact startup environment where your code directly supports next-gen AI deployment.
- Experimentation and self-development are in our culture
- Knowledge sharing and collaboration
- Direct collaboration with top AI researchers and computer vision scientist
- This role is on 2 years employment contract
- Hybrid work setup: ~2–3 days in office per week.