Brian St. Marie

brsm.dev | Malden, MA

Career Overview

Technology executive with 25+ years building and leading consulting teams for enterprise and mid-market clients. Founded and scaled an AI practice, now deploying production AI systems across highly regulated client organizations.

Experience

VP of Technology | iuvo LLC | Boston, MA

Aug 2022 – Present

Refocused iuvo's technology vision toward AI consulting for life sciences, pharmaceutical, and financial services clients.

• Advise executive client leadership and enterprise prospects on AI strategy, ROI modeling, risk mitigation, and compliance frameworks. Deliver hands-on Claude training for small enterprise organizations. Featured as an AI domain expert on the Edge of Excellence podcast.

• Build client POCs and production systems: GBT and statistical modeling, RAG pipelines (pgvector, hybrid lexical/semantic search, reranking), M365 configuration-as-code monitoring (iuvo Guardian), and customized Claude Code deployments (Microsoft Foundry + Entra).

• Secured early acceptance into the Claude Partner Program. Standardized internal AI DevOps and defined the operational baseline for all client engagements.

Service Solutions Director, US | Bulletproof | Brookline, MA

Dec 2021 – Jul 2022

Led Terminal's integration into Bulletproof following acquisition. Managed all client relationships through transition and supervised the expanded Boston team under the new organizational structure. Worked with Bulletproof and GLI leadership on US market strategy and advised on unification of Bulletproof's larger MSP structure into Terminal's management platform.

Director of Technology Solutions | Terminal | Brookline, MA

Nov 2000 – Dec 2021

Led the organization's full transformation from enterprise sales to top-tier managed services provider for regulated businesses. Transitioned from Systems Engineer to IT Services Manager to Director.

• Served as trusted technology advisor for client organizations. Delivered infrastructure migrations for state agencies, municipalities, and healthcare providers subject to HIPAA, IRS Pub 1075, NIST 800-53, PCI, and 201 CMR 17.00.

• Managed ~20 large-scale client projects per year; ~1,500 end users and 2,000 managed devices under management at peak.

• Built and led the engineering team through every phase of company growth, ultimately generating $2.5M ARR with ~100% client retention.

Side Projects

HexN - 234K-parameter heterogeneous GNN, trained by self-play PPO to play classic 4-player Settlers of Catan. Reached #63 of ~50,000 global ranked players on the largest online Catan platform. Wins 83% (vs 25% fair share) against Catanatron AB:2. Trained on HexSet, a purpose-built Catan engine and gym. * Code: https://github.com/0xBrsm/HexN
* Play: https://hexset.brsm.dev

QNN - 184K-parameter transformer encoder + GRU policy modeling human-like behavior in competitive Quake Rocket Arena. Combines offline imitation learning with targeted RL fine-tuning. Achieves p90 expert skill with action distribution in the human band, measured on holdout by NLL and action-mask accuracy.
Code: https://github.com/0xBrsm/QNN
Play: https://qnn.brsm.dev

Cite Unseen (Harvard ALM Thesis) - Plagiarism detection using the public web as corpus. N-gram tokenization, SERP snippet parsing, local sequence alignment scoring. F1 = 0.90 on online content. Dean's Prize for Outstanding ALM Thesis, Honorable Mention.
* Code: https://github.com/0xBrsm/CiteUnseen

Education & Certifications

Master of Liberal Arts, Information Technology - Mathematics & Computation
Harvard University Extension School
Dean's List | Dean's Prize for Outstanding ALM Thesis, Honorable Mention

Bachelor of Science - Biology
University of California, San Diego
UC Regents Scholar

Project Management Professional (PMP) - PMI
Claude Certified Architect - Professional (CCAR-P) - Anthropic
Microsoft Certified Solutions Expert (MCSE) - Cloud Platform and Infrastructure (Charter Member)

Technical

Cloud & Infrastructure: Microsoft Foundry, Azure, Microsoft 365, Fabric, Dataverse, Entra, AWS, GCP, Docker
AI/ML: RAG, MCP, SL/RL, Skills/Plugins, PyTorch, ONNX, llama.cpp, FastFlowLM, ROCm
Coding: C, Go, Java, JavaScript, PowerShell, Python, TypeScript, git, GitHub, CI/CD