I'm Trevor Kellogg — a technology consultant focused on practical AI implementation, agentic workflows, Model Context Protocol integrations, data systems, automation, and digital risk.
$ initialize profile▌
✓ agentic workflows
✓ MCP architecture
✓ enterprise AI governance
✓ SQL-backed intelligence
✓ digital risk & controls
$ status
→ building systems that make AI useful
Selected experience
Working across technology risk, digital controls, and emerging AI implementation questions for financial-services environments. Focused on translating technical systems into clear, defensible operating and control models.
Supported IT SOX and ITGC testing, walkthroughs, internal audit outsourcing, and developed an AI implementation memo covering governance, risk, controls, and ethical considerations.
Built analytical work around dry-cargo and chemical-tanker performance, including TCE analysis, tradelane summaries, and data visualization. Later expanded into practical AI workflows for internal analysis and research.
Current Project
I’m currently implementing AI across a private shipping organization with the goal of making Claude a practical operating layer for employees — not just a chatbot. The work centers on secure connectors, workflow automation, and a custom MCP architecture that brings live internal data into the AI environment.
Custom MCP Architecture
I’m building a custom Model Context Protocol layer that connects Claude to the company’s dynamic SQL database. The goal is to create one centralized intelligence layer where current operational information can be queried naturally without employees manually exporting spreadsheets or copying data between systems.
Company-wide rollout
Deploying Claude across the organization so employees can use AI directly within their day-to-day workflows for research, analysis, drafting, and operational tasks.
Connectors
Implementing connectors that let Claude interact with the tools and information employees already use, reducing repetitive manual work and making AI useful inside existing processes.
Live enterprise context
Replacing copy-and-paste analysis with direct access to structured internal data so AI outputs can reflect current company information instead of stale one-off exports.
Governance
Designing the surrounding access model so employee permissions, sensitive data, and AI capabilities remain controlled as the organization scales its use of Claude.
Projects
Tailgating yard marketplace built with Flask, MySQL, Docker, Jinja, Bootstrap, and FusionCharts. Owned the customer ordering module including search, cancellation, payment, order creation, and manager-facing visualization.
Designed a normalized relational system for a campus used-furniture marketplace with accounts, listings, reservations, payments, reviews, locations, and meeting slots.
Rainbow Six Siege strategy database concept covering operators, weapons, gadgets, maps, bomb sites, site setups, team building, counters, synergies, and meta-aware decision support.
About me
I’m Trevor Kellogg, a 22-year-old from Connecticut and a Virginia Tech graduate. I earned my degree in Business Information Technology with a concentration in Data Support Systems, graduating magna cum laude.
Outside of technology, hockey has been a major part of my life. I played college club hockey at Virginia Tech, an experience that shaped how I think about preparation, teamwork, competition, and performing under pressure.
Professionally, I’m most interested in the point where business, data, and AI meet. I like taking complicated systems, figuring out how information actually moves through them, and building ways for AI and automation to make those systems faster and more useful.
Next