About this role.
This Member of Technical Staff role focuses on data intelligence for large-scale multimodal foundation-model training. The hire will define data-quality standards with researchers, curate and assess datasets, and build automated methods for quality evaluation, domain mixtures, and synthetic-to-real adaptation. They will also own reproducibility, metadata and provenance tracking, CI/CD, and operational tooling for petabyte-scale data and compute pipelines. The position combines applied ML research with production data infrastructure, distributed processing, and cross-functional collaboration in a remote-first startup environment.
Role DNA
A quick view of the complexity, pace, ownership and collaboration implied by the job description.
Job Complexity
5/5Pace & Pressure
5/5Autonomy Level
5/5Communication Load
4/5Salary analysis
Estimated compensation compared with the broader US market for similar roles.
Core skills
Skills and capabilities most closely associated with this opportunity.
Sample interview questions
I would begin by aligning quality criteria to intended model behaviors and downstream evaluations. I would implement measurable checks for integrity, duplication, relevance, distribution coverage, safety, labeling quality, and modality-specific properties, then set acceptance thresholds and monitoring dashboards. Finally, I would validate that the metrics correlate with controlled training outcomes rather than treating proxy metrics as sufficient evidence.
I would version immutable dataset manifests, transformation code, container environments, configurations, and model-training inputs. Each pipeline run would record lineage from raw sources through filtering and sharding to final artifacts, including checksums, metadata, and execution parameters. Orchestration would produce auditable run records so any training or evaluation dataset can be reconstructed or precisely identified.
I would formulate a hypothesis and create controlled experiments that hold model architecture, compute budget, and most training conditions constant. I would compare well-defined mixture ratios against real-only and synthetic-only baselines across representative, held-out evaluations. I would report uncertainty, investigate subgroup regressions, and use ablations to determine whether the observed gain comes from the adaptation method rather than confounding factors.
I would first instrument the pipeline to locate bottlenecks in I/O, serialization, skew, compute utilization, storage layout, and scheduling. I would then improve partitioning and sharding, batch operations, cache reusable artifacts, reduce unnecessary data movement, and select resource configurations based on observed workload characteristics. Changes would be validated with benchmarks that measure both end-to-end throughput and cost per usable training example.
I would clarify the research objective and identify the smallest experiment that can test the underlying assumption. I would propose staged prototypes, estimate quality and infrastructure tradeoffs, and communicate the expected cost, latency, and risks of each option. After selecting an approach, I would deliver measurable milestones and use experiment results to decide whether further scaling is justified.
In this role, you’ll work closely with model researchers, data infrastructure engineers, and cross-functional partners to make sure our data is high quality and can be produced at petabyte scale in a reliable, efficient way. From understanding how data choices show up in model behavior, to building processing pipelines and running the compute behind them, you’ll help ensure our models are trained on the best data we can get.
What you’ll do
Work with model researchers to define what “good data” means for our models, including quality metrics, validation checks, and acceptance thresholds
Explore open source datasets and create internal ones most suitable to build fundamental World Models
Build algorithms for automated data quality assessment, data domain mixtures, and domain adaptation from synthetic to real data.
Track datasets, metadata, provenance, and versions so experiments are reproducible and it’s clear what data went into which training and evaluation runs
Own CI/CD and development tooling for the data stack (GitHub, Python, PyTorch), and automate repetitive workflows to reduce friction
Track and optimize throughput, storage, and compute utilization across pipelines and related assets
What we’re looking for
Strong ML and deep learning fundamentals with experience building and operating large-scale data and/or compute systems
Comfortable moving between research questions and production engineering: you can dig into data, run analyses, and also ship reliable systems
Demonstrated research experience with data compositions, quality, and dataset releases
Ability to design and execute experiments with convincing unbiased outcomes
Practical experience with distributed processing and orchestration (Spark, Ray, Airflow, or equivalents)
Solid Python skills, and familiarity with the tooling around modern model training workflows (datasets, checkpoints, experiment tracking)
Strong instincts around data quality: how to measure it, how to monitor it, and how to prevent regressions as things scale
Able to work in a fast-moving environment, prioritize what matters, and communicate clearly with both researchers and engineers
Bonus: experience with large video datasets, dataset curation for training, or building internal tooling for evaluation/analysis in ML environments
Reka’s Mission
Reka’s mission is to build useful multimodal artificial intelligence and use it to empower organisations and businesses. We are a globally distributed foundation model startup, headquartered in the San Francisco Bay Area, California. Embracing a remote-first approach, our team brings together top talent from around the world. Our founding team, along with many of our team members, has contributed to many of the breakthroughs in AI over the past decade.
Why Reka?
An Elite Team: Collaborate with top-tier engineers, researchers, operators from renowned organizations like Google DeepMind and Facebook AI Research (FAIR) and successful startups, driving innovation in cutting-edge AI technology.
Massive Market Opportunity: Be part of a rapidly growing industry poised to transform multiple sectors globally, offering the chance to make a significant impact.
Mission-Driven Environment: Work alongside a collaborative, mission-focused team dedicated to advancing AI for meaningful applications.
Inclusive and Open Culture: Thrive in an open and inclusive work environment that values diverse perspectives and fosters creativity.
Generous Benefits: Enjoy 5 weeks of paid leave to recharge, comprehensive healthcare benefits including vision and dental, and additional perks that support your well-being.
Visa Support: We provide visa assistance, including H1B and OPT transfers, for US employees to ensure a smooth transition and support your career with us.
Annual salary information is not provided for this position. Explore salary ranges for similar roles in our Salary Directory ›
Upload your resume now
- To unlock remote work opportunities and be discovered by global employers.
This job listing has been manually reviewed by the Jobicy Trust & Safety Team for compliance with our posting guidelines, including verification of the company's legitimacy, accuracy of job details, clarity of remote work policy, and absence of misleading or fraudulent content.
