Rebekah Love

About Rebekah Love

I build applied AI systems that turn complex data into clear, usable insights.Most recently, I designed and deployed an end-to-end AI reporting platform that ingests ~8 million public campaign finance records and generates publishable reports using retrieval-augmented generation (RAG) and structured LLM orchestration.Before this system, producing a single report took 8–12 hours of manual work.Now the same process runs in 10–30 minutes, increasing throughput from roughly 1 report per week to 3–4 reports per day.What makes this work isn’t just using an LLM — it’s designing a system around it: • Structured ingestion and normalization of real-world data • Deterministic analysis layered with controlled AI generation • Context orchestration to ground outputs and reduce hallucinations • Observability, logging, and artifact versioning for reproducibility • Human-in-the-loop workflows for review and quality controlI focus on building production AI systems that are reliable, auditable, and usable by non-technical stakeholders — not just prototypes.I’m particularly interested in greenfield AI systems where architecture, extensibility, and real-world constraints matter.

Experience

1. Founder & Applied Ai Engineer

   Shadow Network Intelligence

   Dec 2024 — Present · Louisville, KY, US

   Built and deployed a production AI platform ingesting ~8 million public campaign finance records into PostgreSQL, enabling scalable analytics and automated report generation through ETL pipelines • Designed and implemented retrieval-augmented generation (RAG) pipelines with structured context assembly and controlled orchestration, improving reliability and reducing hallucination risk through evaluation and prompt engineering • Reduced report generation time from ~8–12 hours manually to 10–30 minutes, increasing throughput from ~1 report per week to 3–4 reports per day • Architected a multi-stage pipeline spanning ingestion, deterministic analysis, context orchestration, LLM generation, and editorial workflow to deliver consistent, publishable outputs for non-technical audiences • Developed containerized backend services (Python, FastAPI, Docker) with API-driven orchestration and AWS S3 artifact storage, enabling reliable and repeatable pipeline execution • Implemented observability, logging, and artifact versioning to ensure reproducibility, auditability, and debuggability of AI-generated outputs • Built a human-in-the-loop editorial workflow allowing non-technical review, revision, and approval of AI-generated content prior to publication • Translated ambiguous reporting and editorial requirements into structured AI workflows and production systems, aligning technical design with real-world usability

Education

• Indiana University Southeast

  Bachelor's degree, Fine and Studio Arts