ERP specialization
I work on ERP software for service and project-based organizations, where planning, operations, finance, and service workflows need to stay understandable across teams.
About
I build business software across ERP systems, FileMaker platforms, web and mobile interfaces, APIs, and practical AI-assisted delivery workflows.
My work sits close to business operations: the planning flows, records, reviews, interfaces, APIs, mobile touchpoints, and automations that teams rely on to keep complex work moving. I care about software that is specific enough to be useful, understandable enough to maintain, and structured enough to improve over time.
Systems work
ERP work needs both technical execution and the judgment to understand where a process can be simplified, automated, or made more reliable for service and project-based work.
Planning workflows
I build and improve planning flows for site, project, and job planning, including scheduling, filtering, load indicators, and multi-job operations.
Operational fit
I look for software decisions that match how people actually plan, review, correct, and repeat their daily work.
FileMaker platforms
FileMaker is most valuable when it is treated as a durable application platform, not only a quick database.
I work with FileMaker data models, scripts, layouts, web viewers, and HTML/CSS/JavaScript interfaces to support custom workflows that need to stay close to business users. The goal is not novelty for its own sake; it is making specialized software easier to use, easier to reason about, and easier to extend safely.
That means paying attention to naming, relationships, script structure, interface details, and the practical maintenance path for the people who will keep using and adapting the system.
Technical range
My ERP and FileMaker work is supported by broader full-stack, mobile, API, and data experience.
Web interfaces
React, HTML, CSS, JavaScript, and FileMaker web viewers for business-facing interfaces that need to stay clear under real operational use.
APIs and data
REST APIs, GraphQL, SQL, JSON schema extraction, and integration work for systems that need reliable handoffs between tools and teams.
Mobile delivery
React Native and TypeScript experience across authentication, REST and GraphQL integrations, frontend implementation, and store release workflows.
AI-assisted workflows
AI is useful when it is applied to specific work: classification, extraction, review, and repeatable support tasks.
I use AI-assisted workflows in practical professional terms: using the OpenAI API to classify email, process invoice and order confirmations, extract structured JSON data, draft implementation options, and support review steps. The emphasis is on making repetitive work more consistent while preserving human responsibility for decisions that require context.
Good AI workflow design starts with the input, the expected output, the failure modes, and the review path. The useful question is not whether a workflow sounds advanced, but whether it reduces friction without hiding risk.
Agentic coding
I use coding agents as structured collaborators for project setup, implementation, review, and verification.
Tools like ChatGPT, Codex, and OpenCode are part of my development workflow. I use them to shape implementation plans, inspect unfamiliar code, draft focused changes, review output, and run verification loops. The value is in disciplined use: clear scope, grounded context, small changes, and evidence before calling work complete.
Agentic coding does not remove engineering judgment. It makes that judgment more important by moving faster through setup and iteration, while still requiring careful review of behavior, security, maintainability, and fit with the surrounding system.
Working principles
A few principles guide how I approach business systems, automation, and delivery work.
Business context first
Technical choices should be grounded in the operational problem, the people using the system, and the cost of getting details wrong.
Durable systems
The best business software is maintainable, traceable, and boring in the right places, especially around critical data and repeated workflows.
Readable implementation
Code, scripts, layouts, and automation should be clear enough to review, change, and support after the first version ships.
Human-reviewed automation
AI assistance is strongest when it handles structure and repetition while people keep responsibility for judgment, quality, and final decisions.