AI Implementation + Automation Developer (Claude/LLM, Field Apps, GitHub)

About us We're a low-voltage and structured cabling contractor based in Memphis, TN. We design and install network cabling, access control, CCTV, and security systems for school districts, universities, general contractors, and commercial clients. Our projects range from a handful of drops in a single building to full campus infrastructure. Over the past year we've built a set of internal AI tools that handle estimating, proposal generation, purchase orders, RFQs, and project handoff packages. They work. They were also built by people whose actual job is running construction projects, not writing software. We want someone who can take what we have, harden it, and help build the next layer. What you'd be working on LLM tooling is the core of it. We use Claude heavily and want someone who understands prompt design, structured outputs, and how to build a tool that produces the same quality result on the hundredth run as it did on the first. Much of our work involves turning messy inputs into clean documents: spec drawings, vendor quotes, bid packages, and spreadsheets that were built by someone in 2014 and never touched since. Beyond that: Automations connecting the systems we already run on: Microsoft 365, Box, Dropbox, QuickBooks, and our CRM Field apps our technicians can use on a phone or tablet for service tickets, install checklists, photo documentation, and closeout paperwork Document generation pipelines producing branded Word, Excel, and PDF deliverables that match our existing templates exactly GitHub structure and discipline, so this stops living in one person's head and starts living in a repo with real documentation Data extraction work: parsing PDFs, drawings, and scanned submittals into usable structured data What we're looking for Hands-on experience with LLM APIs, specifically Anthropic or OpenAI. Not just using ChatGPT, but building on top of it Python and JavaScript. Comfortable with REST APIs, webhooks, and automation platforms like Make, n8n, or Zapier Someone who ships. We'd rather have a rough working tool this week than a polished plan next month GitHub fluency: branches, pull requests, and documentation another developer could pick up cold Clear written English and steady communication. We don't need daily standups, but we do need to know where things stand Helpful but not required: any background in construction, field services, trades, manufacturing, or logistics. Anything where the end user is wearing work gloves and has about fifteen seconds of patience for a bad interface. How we work Practical over elegant. If a tool saves a project manager an hour a week, it's worth building. If it's beautiful and nobody opens it, it isn't. You'll work directly with ownership, so decisions happen in a conversation instead of a committee. We start every engagement with a small paid test project. It's a real problem from our actual workload, scoped to roughly [X] hours, and you get paid for it regardless of whether we continue. That way we both find out how this goes without either of us committing to something blind. If it works, there's steady ongoing work. Several of our internal tools need a permanent owner. Rate and availability: [$XX to $XX]/hour, roughly [X] hours per week to start, with room to grow. Some overlap with US Central time is required for calls, but we don't care where you sit. When you apply: skip the generic cover letter. We'll read past it. Instead, tell us about one automation or AI tool you built end to end. What problem did it solve, who used it, and what did it save in time or money? If you have a repo or a demo we can look at, link it.