Sieve
Member of Technical Staff, Applied Research
San Francisco · Staff+
Sponsorship not specifiedDetected 38 days ago
PythonSwiftMachine LearningDeep LearningPyTorchNLPAR/VRResearchCommunication
Stay score
odds of building a lasting career here
34Risky
Cap-exempt (no lottery)0
Sponsors this role70
Entry-level history0
PERM / green-card track0
Lottery odds40
Fits your clock70
Thin sponsorship signal and lottery-bound. A low-probability bet with your clock running. Prioritize cap-exempt roles and proven entry-level sponsors first.
Lottery odds assume a STEM candidate.
Personalize to your clock →Employer immigration record
from this employer's Department of Labor filings
Files H-1B transfers
5 transfer filings in the last year, covering 5 workers. Median labor-condition decision: 7 days. An employer that already files transfers is one that can take over an existing H-1B.
Sourced from Department of Labor LCA, PERM and prevailing-wage disclosure data. Employer matching is by name, so figures may be split across an employer's legal entities. Absence of a filing means none appears in our copy of the data, not that none exists.
Community outcomes
No reports yet — be the first to help the next applicant.
About the role
- As a Member of Technical Staff, Applied Research at Sieve https://www.sievedata.com/, you'll train and evaluate multimodal models to understand how data shapes their capabilities.
- Your work will span video generation and audiovisual understanding, connecting advances in data curation with measurable improvements in model performance.
- You're likely a good fit if you enjoy moving between research and engineering: reading a paper, implementing a method, debugging a training run, and figuring out whether an apparent improvement holds up.
Responsibilities
- You'll own the research loop end-to-end: identify a model weakness, form a hypothesis about the data or training approach that could address it, build the experiment, and evaluate the results.
- 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.
- Collaborate with research and engineering teams internally and at partner labs to define meaningful problems and communicate results.
Requirements
- Experience designing experiments, establishing baselines, and evaluating results critically.
- Familiarity with modern generative or multimodal architectures, such as diffusion models or transformers.
- Strong communication skills and the ability to explain findings, tradeoffs, and uncertainty clearly.
- Experience training video, image, or audio generation models.
- Experience with distributed training and GPU performance optimization.
- Experience as an early hire at a startup.
Benefits
- 401k + Full Health Insurance
Company info
- Sieve is a multi-modal lab curating the world's highest-quality training datasets - spanning video, audio, images, text, and 3D. We combine exabyte-scale data infrastructure and novel multimodal understanding techniques that push the frontier of foundation models. Video alone makes up 80% of internet traffic, and across modalities, data has become the enabling medium powering creativity, communication, gaming, AR/VR, and robotics. Sieve exists to solve the biggest bottleneck in the growth of these applications: high-quality training data.
- We partner with top AI labs and did $XXM last quarter alone, as a team of ~30 people. We also raised our Series A from Tier 1 firms such as Matrix Partners https://matrix.vc/, Swift Ventures https://www.swift.vc/, Y Combinator https://www.ycombinator.com/, and AI Grant https://aigrant.com/.
- Sieve combines access to diverse multimodal data, infrastructure to process it at scale, and close relationships with the teams building frontier models. This gives us a unique opportunity to study what makes training data effective-and turn those findings into better models and datasets.
- You'll join a small team building our research capabilities, with ownership over experiments, training systems, and the decisions those results inform.
- Sieve is a multi-modal lab curating the world's highest-quality training datasets - spanning video, audio, images, text, and 3D.
- We combine exabyte-scale data infrastructure and novel multimodal understanding techniques that push the frontier of foundation models.
- Video alone makes up 80% of internet traffic, and across modalities, data has become the enabling medium powering creativity, communication, gaming, AR/VR, and robotics.
- Sieve exists to solve the biggest bottleneck in the growth of these applications: high-quality training data.
- We partner with top AI labs and did $XXM last quarter alone, as a team of ~30 people.
- We also raised our Series A from Tier 1 firms such as Matrix Partners https://matrix.vc/, Swift Ventures https://www.swift.vc/, Y Combinator https://www.ycombinator.com/, and AI Grant https://aigrant.com/.
- Sieve combines access to diverse multimodal data, infrastructure to process it at scale, and close relationships with the teams building frontier models.
- This gives us a unique opportunity to study what makes training data effective-and turn those findings into better models and datasets.
- Sieve exists to solve the biggest bottleneck in the growth of these applications:
- high-quality training data.
This listing is sourced directly from Sieve's careers page and normalized into a canonical job model.