Where Physical Reality Collides with Software

Much of the modern technology industry designs systems for sterile, idealized environments—sandboxes where networks never drop, databases never contain dirty regulatory liabilities, and physical inventories behave predictably like database rows.

“Most tech workers solve imaginary problems in pristine code; I solve messy, high-stakes operational problems where physical and business realities collide. I harness agentic engineering and multimodal AI not to generate boilerplate, but to automate real-world workflows that move numbers.” — Allen Christopher Tools, Operating Philosophy

I work in agentic coding workflows and build with multimodal AI models (Gemini) daily. Rather than producing generic wrappers or superficial prototypes, I deploy agentic AI as an operational force multiplier:

• Physical Inventory Audits: Visual shelf scans that inspect walk-in coolers without manual clipboard friction.
• Yield Reconciliation: Balancing physical smokehouse shrinkage against live POS transactions in real time.
• Margin Protection: Stopping silent balance sheet leakage where real-world kitchen operations and digital ledgers drift.
• Regulatory Governance: Enforcing automated PII and compliance audits across enterprise data lakes.

An Operational Crucible: From the Line to the Enterprise

Domain mastery is rarely forged on a straight, corporate ladder. It comes from having skin in the game across operations, risk, governance, and code.

Yellow Rose IMS: Physical Yields Meet Edge AI Vision

Barbecue is a brutal business of razor-thin margins and physical physics. Raw beef brisket shrinks between 44% and 52% during a 14-hour smoke cycle. Standard inventory management software assumes a 1:1 recipe conversion, leaving operators blind to true meat yields, food costs, and shrinkage.

Rather than paying thousands of dollars for clunky enterprise restaurant SaaS that kitchen staff refuse to use, I designed and built the Yellow Rose Inventory Management System from the ground up.

Key Engineering Breakthroughs

• Multimodal Shelf Vision: Kitchen staff capture a walk-in storage photo. The image routes via a lightweight Cloudflare Worker to gemini-1.5-flash, returning structured JSON counts of briskets, pork butts, and soda bibs—eliminating manual clipboard counts.
• Real-Time Yield Reconciliation: A custom Python depletion engine compares Clover POS register receipts against raw butcher weights and physical cook yields, isolating kitchen trimming loss, smoker shrinkage, and unaccounted variance.
• Two-Way Edge Sync: High-speed local SQLite caching paired with bi-directional Google Sheets API synchronization delivers live cost-of-goods figures to floor managers without software training overhead.

Technical Domain Range

A consolidated view of the tools and systems applied across physical operations, enterprise databases, and cloud runtimes:

Starting a Conversation

I am energized by high-leverage challenges at the intersection of business operations, data systems, and agentic workflows.

Whether you need someone to diagnose where your operational data is leaking margin, build custom edge automation that connects physical workflows to the cloud, or architect systems that outlast off-the-shelf software—let's speak directly.