Lauren Banawa

I'm an ML engineering leader with hands-on experience building LLM pipelines, document extraction systems, and agentic tooling, in addition to leading the teams that ship them.

Experience
··
  • ML lead for JARVIS, Imagineering's assistant — 20M+ files, >80% adoption (~10k DAU)
  • Built ML for document extraction — YOLO + LLM/VLM prompts, Pydantic validation
  • Cut extraction costs from ~$2M to under $100k per release via self-hosted GenAI
  • Built Evaluation-Driven Development — team of 10+, golden datasets, benchmarks
  • Partnered with PMs and stakeholders to prioritize ML features against org SLAs
  • Orchestrated Lexis+ with Protégé delivery — 4 capabilities across US, APAC, EMEA
  • Managed 8 engineers as Lead DS — hiring, development plans, promotion sponsorship
  • Managed annual budget across 5 projects via custom OPEX/CAPEX dashboards
  • Raised test coverage 20% → 95%, cut cycle time ~20 days → 12 via AI-native workflows
  • Delivered NLP/LLM pipelines for NER, NLU, and Q&A in finance & legal
  • Built ML for AIVA, mortgage automation for the top 10 banks (AWS re:Invent 2022)
  • Owned the ML lifecycle — EDA, annotation, training, deployment, retraining
  • Built document classification at >90% accuracy (BERT, RoBERTa, BigBird) on AWS
  • Mentored junior engineers, set best practices for experiment reproducibility
  • Ran business operations and P&L — client acquisition to contractor management
  • Led soft goods design and production, concept through small-batch manufacturing
  • Built partnerships with the Carolina Textile District and Mosaic Changemakers
  • Led design quality reviews with third-party production partners
  • Prototyped designs via computer-assisted patterning (CAD, CLO)
  • Wrote tech packs enabling scalable garment production, domestic and global
  • Self-Maintaining Hierarchical Tool Routing System for LLM Agent Ecosystems 2026
  • Context-Restricted Prompt Routing for Structured Multimodal Information Extraction 2026
Upcoming
  • Context-Restricted Prompt Routing for Structured Multimodal Information Extraction 2026
  • Hierarchy-Aware Hard Negative Mining for Document Retrieval 2026
Filed
  • Self-Maintaining Hierarchical Tool Routing System for LLM Agent Ecosystems 2026
  • Gated Multimodal Mixture of Experts for Document Classification 2025
Education

B.A. Theatre Production, University of South Florida

Fun
Home

Tampa · Pensacola · Orlando · Johnson City · Los Angeles · Valencia (ES)

Elsewhere
LinkedIn