Actively recruiting / 35 applicants
We’re here to help you
Juliana Torrisi is in direct contact with the company and can answer any questions you may have. Email
Juliana Torrisi, RecruiterUnmuted is building the intelligence layer for political campaigns — AI agents that handle opposition research, voter targeting, donor prospecting, and fundraising automation.
This role owns the models, pipelines, and data infrastructure that make the agents actually work. You'll report to Matthew (CTO) and operate within his technical architecture.
The Role
You'll build the ML systems that power Unmuted's five AI agents — from donor scoring models and voter targeting to fundraising email generation and opposition intelligence. The work spans data engineering, model development, and agent pipeline design. You'll work closely with Matthew on architecture and independently execute on well-scoped sprints.
This is not a research role. The emphasis is on systems that work reliably in production for real campaigns with real deadlines. Speed and accuracy matter equally.
What You'll Build
First 6 weeks
- Donor scoring model — capacity and motivation axes from L2 + upload data
- L2 voter file ingestion and enrichment pipeline
- Fundraising email corpus collection and training pipeline
- Voter segmentation engine with natural language query interface
Weeks 7–12
- Scenario simulation engine — turnout projections, win probability, coalition shifts
- A/B message testing and persuasion elasticity modeling
- Outcome tracking infrastructure across campaign cycles
- Operator knowledge integration — annotating agent outputs with domain expertise
What We're Looking For
Required
- 4+ years in ML engineering or applied AI — models shipped to production
- Strong Python, data pipelines, and model evaluation
- Experience with LLM integration — prompt engineering, fine-tuning, or RAG
- Comfortable with messy, real-world datasets (not just Kaggle-clean data)
- Fast, pragmatic builder — you know when good enough ships and when it
Strong Plus
- Experience with voter file data, census data, or behavioral scoring
- NLP work — classification, entity extraction, sentiment analysis
- Multi-agent system design
- Political technology, civic tech, or public sector data experience