AI Strategy and Automation Engineer
LegalMatch
Reno, NV
May 2025 – Present
- Engineered a multi-stage LLM workflow (Python) to audit ~7,000 law-library articles; flagged outdated content on ~10% of pages and saved an estimated 13+ months of manual review time.
- Built LLM evaluation and regression-test harnesses to compare models, prompts, settings, cost, and output quality; diagnosed and resolved a production tone defect via prompt redesign.
- Optimized a SageMaker embedding + clustering pipeline (DistilRoBERTa + K-means) by excluding empty/irrelevant text during feature generation, improving runtime 50x+ while preserving results.
- Productionized AWS-scheduled monitoring (ECS/EventBridge) for site quality and SEO signals; exported metrics to Datadog alerts and delivered an on-demand page-speed validation tool.
- Built a lead-identification pipeline enriching ~1M visitor IPs with firm detection and non-client filtering; delivered real-time sales notifications and recommended vendor adoption after benchmarking.
- Built an LLM-guided intake MVP (Python + OpenAI API) and began extension to voice AI intake using playbook-driven dialog and tool-integrated data capture.