shoults.ai

AI executive & hands-on builder · NWA (Northwest Arkansas)

Catherine Shoults, PhD MPH

I turn ambiguous business goals into AI platforms that ship.

AI executive and hands-on builder-leader with 15 years across data-intensive, regulated industries. PhD-trained in biomedical informatics, shipping production LLM and agentic systems on AWS — and building this site in the open to show how.

Get in touchSee my experience

knowledge graphWhat I work in

AI engineering
  • AI Engineering — PhD-trained in biomedical informatics (NLP). I build production LLM and agentic systems end to end — and set the strategy for where they go.
  • LLMs & GenAI — Productionizing frontier and open-weight models — OpenAI, Claude, and local Qwen — routed by task for cost and data privacy.
  • Agentic systems — Multi-agent orchestration that ingests unstructured input, extracts entities, enriches against criteria, and generates ranked, sourced deliverables.
  • RAG architectures — Retrieval-augmented generation over large enterprise corpora, with citation discipline and hallucination guardrails.
  • Model Context Protocol — MCP-exposed tool and knowledge layers that let agents query enterprise data in natural language.
  • Vector databases — Embedding models and vector search as the retrieval backbone of RAG systems.
  • Prompt engineering — Structured prompt systems with per-section budgets, output contracts, and linter-enforced sourcing rules.
  • LLM evaluation & cost — Systematic model evaluation, statistical comparison across runs, and cost optimization via task-appropriate model routing.
  • NLP — Pre-generative transformer NLP and weak supervision (SpaCy, Prodigy) — the pharmacovigilance pipeline at the center of my PhD.
Cloud architecture
  • Cloud Architecture — End-to-end AWS platforms for regulated, data-intensive work — including the serverless architecture serving this very site.
  • Bedrock & AgentCore — Managed foundation models and agent runtimes as the compute layer for production GenAI pipelines.
  • Serverless — Event-driven architectures on Lambda, ECS, and EventBridge that scale to zero and bill like it.
  • AWS data stores — DynamoDB, Neptune, RDS, and S3 chosen by access pattern — from entity stores to graph to object corpora.
  • Snowflake — Cloud data warehousing and agentic querying for enrichment and decision science.
  • Infrastructure as Code — Typed, tested, reviewed infrastructure. This site's AWS stack is public CDK.
  • Knowledge graphs — Graph databases over 40k+ document corpora, exposed to agents through an MCP retrieval layer.
  • Data engineering — ETL/ELT with data auditing, entity extraction, and large-scale corpus processing pipelines.
Stats & epidemiology
  • Stats & Epidemiology — MPH in epidemiology (Yale). Fifteen years of causal inference and statistical rigor underneath every model I ship.
  • Survival analysis — Time-to-event modeling for efficacy, safety, and outcome prediction.
  • Time-series forecasting — Point-in-time, leakage-aware forecasting — market downturn predictors and literature-driven signal detection.
  • Causal inference — Propensity score matching and study design for acquisition prediction and drug efficacy/safety.
  • Boosted ML & trees — Gradient-boosted models and decision trees with leakage-aware feature engineering.
  • Clinical trials & regulatory — Phase I–first-in-human trials, statistical analysis plans, and FDA-aligned protocols across drug development.
Cross-cutting practices
  • Responsible & ethical AI — Data governance, model evaluation standards, and AI ethics — the subject of my keynotes and university lectures.
  • Executive & board communication — Translating complex AI into executive-ready insight for boards, executives, and non-technical teams.
  • Team leadership — Building and leading multi-national (US/UK/India) cross-functional teams of engineers, scientists, and statisticians.
  • AI strategy & roadmaps — Defining and executing enterprise AI roadmaps — turning ambiguous business goals into shipped platforms.
  • Keynote speaking — A frequent keynote speaker and AI evangelist on the practical, ethical application of AI.

experienceRecent work

VP, AI & Analytics

SymBiosis Capital Management · 2023-04–now

Define and execute the firm-wide AI and data strategy, embedding GenAI, agentic systems, and cloud-native analytics across the entire venture lifecycle — sourcing, due diligence, and portfolio support.

Research Scientist

University of Arkansas for Medical Sciences · 2018-08–2023-08

Engineered a scalable pharmacovigilance pipeline (PhD dissertation) mining massive unstructured social-media corpora with transformer models and weak-supervision frameworks (Prodigy, SpaCy) — architecting custom programmatic labeling and fine-tuning workflows before the commercial LLM paradigm.

Director of Drug Development

Orbis Biosciences · 2015-07–2018-06

Led cross-functional teams of researchers, engineers, statisticians, clinicians, and regulatory experts through complex drug-development programs, including animal trials and first-in-human clinical trials.

Full experience →

speakingHave me speak

I'm a frequent keynote speaker and AI evangelist — briefing executives, boards, clinicians, and educators on the practical, ethical application of AI. Recent keynotes span hospice and palliative care, healthcare, and public education.

Book a talk