Software Engineer, Research - Human Data
Build full-stack systems, tools, and infrastructure for human feedback collection, AI model alignment, and evaluation. Collaborate with researchers to scale production systems and enhance model safety in a fast-paced environment.
About the job
In this role, you will:
- Build and maintain robust full-stack systems for feedback collection, data labeling, and evaluation pipelines, while maintaining high levels of security.
- Translate experimental alignment research into scalable production infrastructure, including inference and model training stacks.
- Design and iterate on user-facing tools and backend services to support high-quality data workflows.
- Partner with researchers, engineers, and program leads to shape feedback loops and model interaction paradigms.
- Drive infrastructure improvements that enable faster iteration and scaling across OpenAI’s frontier models, from internal research tooling all the way to production ChatGPT.
You might thrive in this role if you:
- Have strong software engineering fundamentals and experience building production systems at scale.
- Enjoy full-stack development with end-to-end ownership — from backend pipelines to user interfaces.
- Are motivated by high-impact collaboration with research teams and solving novel, ambiguous problems.
- Are excited to shape how AI systems learn from human preferences and reflect a broad range of human values.
- Care deeply about inclusive tooling and building systems that enhance model safety, reliability, and usefulness.
Skills
Full-Stack Development, Backend Pipelines, User Interfaces, Data Pipelines, Model Training, Inference, Kubernetes, Python, React, Postgres
Similar jobs
Data Engineering jobs
Bevi
Boston, MA
Build and maintain dbt models, Snowflake semantic layers, and ingestion pipelines across business functions while improving data quality and resilience. The role requires 4–6 years of analytics or data engineering experience, strong dbt and SQL expertise, and a quantitative bachelor's degree.
Stuut
San Francisco, CABuild and own Stuut’s foundational data platform, including ingestion pipelines, canonical models, semantic layers, and observability. The role requires 3+ years of production data pipeline experience with Python, SQL, cloud warehouses, and ETL/ELT tooling.
Ontic
United States
Build scalable analytics engineering infrastructure, SaaS data models, and AI-enabled workflows that support enterprise decision-making. The role requires 3–6 years of hands-on analytics or data engineering experience, strong SQL and modern data modeling expertise, and cloud data warehouse experience.
Underdog Fantasy
United States
Build and operate the data platform supporting automated regulatory reporting for a prediction markets business. The role combines SQL and dbt development, end-to-end data investigation, automated validation, and cross-functional ownership under strict deadlines.
Attentive
United States
Build and operate distributed data applications powering large-scale audience segmentation and real-time personalization. The role requires 2–4 years of software engineering experience, backend development skills, and familiarity with databases, algorithms, and distributed systems.