About us:
Blue Rose Research builds data and AI tools that help Democrats win elections. Our team combines engineering, data science, and political strategy to power decisions for the country’s top campaigns and progressive organizations. We forecast elections, test ads, and use generative AI to help campaigns understand what’s happening in the news; then respond fast with messages that actually work. We have guided how hundreds of millions of dollars are spent in modern campaigns. We’re a small, mission-driven team that builds fast, experiments boldly, and helps progressives communicate and win—guided by curiosity, purpose, and a genuine desire to use technology for good.
The RoleWe’re seeking a mission-driven software engineer who thrives at the intersection of AI systems and statistical modeling. You’ll help design and scale tools that use generative AI and LLM tools to power our causal inference models and predictive analytics—giving progressive campaigns sharper insight and faster response.
As a member of our Message Testing Data Science Team, you’ll build products that connect modern LLM applications (like news summarization, content generation, and evaluation agents) with robust hierarchical modeling pipelines that explain why messages work and who they persuade.
If you love both writing efficient production code and interpreting complex statistical models—and want your skills to make a real-world impact on elections—this role is for you.
ResponsibilitiesAI & LLM Product Development
- Develop and scale LLM-based tools that analyze news, social media, and message performance, combining hierarchical Bayesian models with cutting-edge LLM tooling.
- Build content-generation and summarization systems that interface with campaign data and modeling outputs.
- Automate and optimize data labeling, model fine-tuning, and model evaluation loops using Python-based frameworks.
Statistical Modeling & Causal Inference
- Design, train, and interpret large logistic regression and hierarchical models to estimate causal effects of political messages and ads.
- Build pipelines that integrate experimental and observational data for treatment effect estimation.
- Communicate results and uncertainty clearly to both technical and strategic audiences.
- Collaborate on A/B testing frameworks and survey experiment design.
Core Engineering & Infrastructure
- Architect and maintain modeling and AI infrastructure with clear separation of concerns, reproducibility, and scalability.
- Write clean, efficient, and well-tested Python code; manage data flow and model deployment.
- Develop internal APIs and lightweight web tools for data visualization, LLM interactions, and model monitoring.
Data Engineering & Political Data
- Build and maintain ETL workflows, SQL-based schemas, and workflow automation.
- Integrate voter files, survey data, and ad-performance metrics into unified modeling pipelines.
- Support data reliability, performance monitoring, and quality assurance.
- Proficient in Python and comfortable writing and querying SQL (BigQuery).
- Skilled in statistical modeling (e.g., logistic regression, hierarchical/multilevel models, causal inference, treatment effects).
- Familiar with AI/ML frameworks (e.g., OpenAI, Hugging Face, LangChain) and data science libraries (e.g., scikit-learn, statsmodels, PyMC, or Stan).
- Comfortable working with large datasets and experimental data pipelines.
- Clear communicator: able to translate technical results into actionable insights.
- Collaborative, curious, and impact-driven—you want your code to make a difference.
Bonus Points:
- Experience with survey research, political data, or voter files.
- Familiarity with cloud infrastructure (GCP, AWS) or orchestration tools (Airflow, Prefect).
- Experience building evaluation systems for LLMs or integrating generative models into production workflows.
- Salary: $140,000 – $190,000 annually, commensurate with experience.
- Benefits: Competitive health, dental, and vision coverage; generous leave; and a supportive, mission-driven culture.
- Work setup: Fully remote team with an NYC office and regular in-person meetups (NYC & DC). Most of our work happens on East Coast time.
