Multimodal ML Applied Research Engineer — Onsite SF

Sieve; San Francisco (CA); On-site; USD 180,000 - 280,000; Full time

Benefits offered by this job

401k + Full Health Insurance

Breakfast, Lunch, and Dinner covered

Ubers covered home

Job summary

Sieve is seeking a Member of Technical Staff, Applied Research, in San Francisco to train and evaluate multimodal models. You will connect data curation with model performance, own the research loop from hypothesis to evaluation, and build reproducible training pipelines for video generation and multimodal tasks.

You will bridge research and engineering, implement methods from papers, debug training runs, and drive reliable systems for sourcing, curating, and improving training data.

Qualifications

• 2+ years of experience in machine learning research or engineering.
• Strong Python and PyTorch skills, including the ability to implement, debug, and modify model training code.
• Experience designing experiments, establishing baselines, and evaluating results critically.
• Familiarity with modern generative or multimodal architectures, such as diffusion models or transformers.
• Comfortable working with large datasets and diagnosing training bottlenecks, instability, and data quality issues.
• Able to turn ambiguous research questions into concrete experiments and maintainable systems.
• Strong communication skills and the ability to explain findings, tradeoffs, and uncertainty clearly.

Responsibilities

• Train and post-train models for video generation and multimodal understanding.
• Design controlled experiments to measure how data selection, mixtures, and supervision affect model capabilities.
• Build evaluations that reveal specific model weaknesses, using quantitative metrics and human judgment.
• Develop reliable training infrastructure, including distributed training, efficient data loading, checkpointing, and experiment tracking.
• Turn research findings into improvements in data curation pipelines and products.
• Collaborate with research and engineering teams internally and at partner labs to define meaningful problems and communicate results.

Skills

Python

PyTorch

Experiment design

Diffusion models

Multimodal ML

Large datasets

Communication

Job description

Sieve is seeking a Member of Technical Staff, Applied Research, in San Francisco to train and evaluate multimodal models. You will connect data curation with model performance, own the research loop from hypothesis to evaluation, and build reproducible training pipelines for video generation and multimodal tasks.

You will bridge research and engineering, implement methods from papers, debug training runs, and drive reliable systems for sourcing, curating, and improving training data.