Machine Learning Engineer Data AI Training

@ Canva

About the Role

Join Canva as a Machine Learning Engineer Data AI Training where you will be responsible for high-performance data acquisition, processing, and annotation to enable cutting-edge model training. Work alongside research teams to drive novel model development and evaluation with petabyte-scale data.

Key Focus Areas

• Data Acquisition: Develop scalable tools and pipelines from diverse sources.
• Curation: Engineer robust solutions for deduplication, quality assessment, and filtering data.
• Data Infrastructure: Build high-throughput tools for querying large-scale data pools.

Primary Responsibilities

• Collaborate with research teams to ensure quality data flow.
• Curate targeted datasets using clustering, embedding similarity, and automated scoring.
• Extract and communicate insights regarding dataset composition and biases.
• Build performant, parallel algorithms optimized for throughput and cost-efficiency.
• Engineer intuitive interfaces to enable researchers to sample and interact with large datasets.
• Work with paired multimodal data ensuring alignment and synchronization.
• Utilize high-performance computing frameworks (Ray, Spark, torch.distributed, DeepSpeed) with cloud infrastructure.

Additional Role Insights

This role extends into designing scalable pipelines and building interfaces for data exploration and human feedback integration. It includes responsibilities such as statistical analysis, synthetic data generation, and rigorous evaluation to support generative AI models at Canva.

Key Skills/Competency

• Data Acquisition
• Data Curation
• Parallel Computing
• ML Frameworks
• Python
• Statistical Analysis
• Cloud Infrastructure
• Data Visualization
• Automation
• Multimodal Data