Member of Technical Staff - Data Quality Engineer (Post-training)
San Francisco, CA • FullTime
Posted 6mo ago
Remote Work Policy
On-site
Employment Type
FullTime
Categories
Applied AI Engineer
About the job
Reflection is a research lab dedicated to making intelligence open and accessible. As a member of the Data Team, you will ensure the data used to train and evaluate our models meets high standards for quality, reliability, and impact. You will work closely with world-class researchers on post-training teams to translate abstract notions of "good data" into concrete, measurable standards that scale across large data campaigns. This role requires engineers who combine strong engineering fundamentals with a deep curiosity about data quality and its impact on model behavior.
Responsibilities
- Own upstream data quality for LLM post-training and evaluation by analyzing expert-developed datasets and operationalizing quality standards for reasoning, alignment, and agentic use cases.
- Partner closely with research and post-training teams to translate requirements into measurable quality signals, and provide actionable feedback to external data vendors.
- Design, validate, and scale automated QA methods, including LLM-as-a-Judge frameworks, to reliably measure data quality across large campaigns.
- Build reusable QA pipelines that reliably deliver high-quality data to post-training teams for model training and evaluation.
- Monitor and report on data quality over time, driving continuous iteration on quality standards, processes, and acceptance criteria.
Requirements
- Strong engineering fundamentals with experience building data pipelines, QA systems, or evaluation workflows for post-training data and agentic environments.
- Detail-oriented with an analytical mindset, able to identify failure modes, inconsistencies, and subtle issues that affect data quality.
- Solid understanding of how data quality impacts training (SFT and RL) and evaluation, with the ability to translate quality concerns into concrete signals, decisions, and feedback.
- Experience designing and validating automated quality checks, including rule-based systems, statistical methods, or model-assisted approaches such as LLM-as-a-Judge.
- Comfortable working autonomously, owning problems end-to-end, and collaborating effectively with researchers, engineers, and operations partners.
Benefits
- Top-tier compensation: Salary and equity structured to recognize and retain our talent globally.
- Stock options
- Comprehensive medical, dental, vision, and life insurance
- Annual wellness allowance
- Lunch and dinner provided in the office daily
- 22 weeks paid parental leave for all new birthing and non-birthing parents
- Unlimited paid time off in the U.S.
- 30 days paid time off in the U.K.
- Sponsorship support for visas and long-term immigration pathways
- Regular off-sites, happy hours, and team celebrations