I build production GenAI systems for regulated environments: RAG, agents, evaluation, observability, security controls, data architecture, and AI governance.
Areas of focus: Applied AI architecture · GenAI solutions · LLM evaluation · AI governance engineering · regulated AI deployment · public-sector AI
Dublin-based Applied AI and GenAI systems leader. Experience across Accenture, NSF, federal AI delivery, and production GenAI systems. Focus areas: regulated GenAI, LLM evaluation, RAG and agents, AI governance, and applied AI architecture.
A build-versus-buy decision brief and cost model for AI inference after the flat-rate era. The real hedge against token metering is switchability, an OpenAI-compatible interface against interchangeable open-weight providers with a frontier model in reserve, not owning GPUs. Managed open-model inference usually wins on cost; owned hardware only pays past roughly 100B tokens/month with a platform team.
White Paper · AI Infrastructure · LLMOps · Cost Modeling · Build vs Buy
Built greenfield AI governance, technical-review, security, community, and data-architecture foundations inside NSF's newly formed Chief AI Officer function. Owned AI use-case intake improvements, created risk-classification and technical-review patterns, created an AI deployment playbook, co-chaired a 100+ member AI Community of Practice, served as a voting engineering review board member for AI systems, ran AI security tabletop exercises, supported Microsoft Copilot rollout, architected a production vector/graph capability for research-impact intelligence, and liaised with NASA, DOE, NAIRR stakeholders, and other federal partners.
AI Governance · Responsible AI · Public Sector AI · Technical Review · Vector + Graph Architecture · AI Security Table-Tops · Interagency Coordination
Production-grade ML cost optimization: model selection, batching, caching, cold-start mitigation, inference routing, and observability — applied to live LLM-backed workloads under regulated SLAs.
LLMOps · Cost Optimization · Inference Routing · Observability · Production AI
Public model-readiness benchmark for regulated enterprise AI, evaluating capability, reliability, governance behavior, groundedness, security reasoning, cost, and latency.
LLM Evaluation · Model Readiness · RAG Groundedness · AI Governance · Cost / Latency
Evaluation research comparing microagent atomic policy checks with monolithic auditing across 193 service contracts and 7,913 labeled policy checks.
Contract Compliance · Microagents · Token-Level Calibration · AI Governance
Applied AI architect with delivery experience across Accenture (commercial), NSF (federal AI oversight, GS-15 Lead Data Scientist), and independent production GenAI systems. Focus areas: regulated enterprise AI deployment, RAG and agentic systems, LLM evaluation pipelines, AI observability, prompt-injection and tool-use security, and AI governance operating models. Based in Dublin, Ireland.