Machine Learning Engineer

CANTINA RESEARCH SINGAPORE PTE. LTD.; Singapore; On-site; SGD 120,000 - 180,000; Full time

Benefits offered by this job

Health insurance

Global travel insurance

Monthly stipend

Home office equipment

Job summary

Cantina Labs, a social AI company, seeks an ML Engineer to build and scale data pipelines for video and multimodal datasets used in model training. You will own end-to-end data processing from ingestion to training-ready datasets with a focus on speed, reliability, and cost-efficiency.

Collaborate with curation and modeling teams to operationalize evolving dataset recipes and improve model outcomes through scalable, containerized pipelines and cloud optimizations.

Qualifications

• Experience building or scaling large-scale data systems and pipelines for ML.
• Experience with distributed data processing frameworks (e.g., PySpark, Ray).
• Hands-on with container-based pipelines (Docker, Kubernetes).
• Experience with cloud storage/compute (AWS, GCS, Azure).
• Experience with VLM-based captioning pipelines or quality scoring models.
• Familiarity with CLIP-based or embedding-based filtering and semantic data selection.

Responsibilities

• Design and scale distributed data pipelines for preprocessing, dataset generation, and refreshes.
• Own workflow orchestration, job scheduling, monitoring, and failure recovery for large-scale data jobs.
• Implement and maintain containerized pipeline infrastructure using Kubernetes or equivalent.
• Optimize cloud storage and data movement across providers for cost and throughput.
• Define dataset storage layouts, versioning, caching, and access patterns.
• Develop curation pipelines to select, filter, and retain training data.
• Build and improve VLM-based captioning and metadata generation workflows.
• Develop quality and aesthetic scoring models and deduplication pipelines at scale.
• Analyze dataset composition to improve training outcomes and standards for data quality.

Skills

Data pipelines

PySpark

Airflow

Docker

Kubernetes

Python

Cloud platforms

VLM/CAPTIONING

CLIP-based filtering

Tools

Docker

Kubernetes

PyTorch

FFmpeg/OpenCV

PySpark/Ray

Airflow

Job description

About Cantina:

Cantina Labs is a social AI company, developing a suite of advanced real-time models that push the boundaries of expression, personality, and realism. We bring characters to life, transforming how people tell stories, connect, and create. We build and power ecosystems. Cantina, our flagship social AI platform, is just the beginning.

About the Role:

Cantina is expanding, and we're looking for an ML Engineer to join our growing Singapore team! In this role, you will build and scale systems for ingesting, processing, and delivering large-scale video and multimodal data for model training. You'll own the full pipeline — from raw content to curated, filtered, and training-ready datasets — with a focus on speed, reliability, reproducibility, and cost-efficiency. You'll partner closely with curation and modeling teams to operationalize evolving dataset recipes and iterate on approaches that improve model outcomes.

What You’ll Do:

• Design and scale distributed data pipelines for preprocessing, dataset generation, and repeated dataset refreshes
• Own workflow orchestration, job scheduling, monitoring, and failure recovery for large-scale data processing jobs
• Implement and maintain containerized pipeline infrastructure using Kubernetes or equivalent orchestration systems
• Optimize cloud-based data storage and movement across providers (AWS, GCS, or Azure) for cost, throughput, and operational efficiency
• Define and implement best practices for dataset storage layout, versioning, caching, retention, and access patterns
• Design and implement curation pipelines that determine which video and image content is selected, filtered, and retained for model training, including image-text pair datasets used in joint training regimes
• Build and improve VLM-based captioning and metadata generation workflows at scale across both video and image data
• Develop and apply quality and aesthetic scoring models, CLIP-based semantic filtering, and other signal-extraction approaches for data selection
• Build tooling to support deduplication workflows at scale, including near-dedup and exact deduplication pipelines over large video corpora
• Analyze dataset composition, identify quality issues, and iterate on curation logic to improve training outcomes
• Define and evolve standards for what constitutes high-quality, training-ready video data across different training regimes

What You’ll Bring:

• Strong hands-on experience building or scaling large-scale data systems and pipelines for machine learning, including dataset curation, filtering, and quality improvement
• Experience with distributed data processing frameworks such as PySpark or Ray, and orchestration tools such as Airflow or equivalent
• Familiarity with containerization and container orchestration, including Docker and Kubernetes
• Experience working with cloud-based data storage and compute (AWS, GCS, and/or Azure), including tradeoffs around cost, throughput, storage layout, and access patterns
• Experience with VLM-based captioning pipelines or quality/aesthetic scoring models for video or image data, including curation of image-text pair datasets for joint image-video training
• Familiarity with CLIP-based or embedding-based filtering and semantic data selection techniques
• Familiarity with video and media processing tools such as FFmpeg, PyAV, DALI, or OpenCV, and relevant libraries such as Decord, torchvision, PyTorchVideo, or torchaudio
• Proficiency in Python
• Strong problem-solving, communication, and documentation skills

Benefits We Offer:

• Competitive salary and generous company equity
• Personal time off and paid holidays
• Health insurance
• Global travel insurance: Covers you when traveling internationally
• Monthly spending stipend: $500 (~S$635)
• Equipment: All equipment needed for your home office