Part-Time AI Agent Engineer

Part-Time AI Agent Engineer (OpenClaw / Python) Automating Operations for a U.S. Sign Manufacturing Company We are a U.S.-based electric sign manufacturing and installation company serving national retail and restaurant brands across North America. Our work includes design, fabrication, permitting, logistics, and installation of signage for multi-location clients. We are looking for a practical AI engineer to help us design and build a set of production-ready AI agents that automate parts of our operational workflow. This is not a research project or demo. We are looking for someone who enjoys building systems that solve real operational problems for small teams. Our goal is to develop AI agents that reduce repetitive administrative work so our project managers can focus on higher-value tasks. We expect the system to evolve over time, so this role may become an ongoing part-time collaboration if the first phase goes well. You will have significant input into system architecture and how the platform evolves over time. What You’ll Build You will design and implement one or more AI agents (or a small multi-agent system) that integrate with our existing tools and assist with operational workflows. We are currently exploring frameworks such as OpenClaw, but we are open to similar agent frameworks if you believe they are better suited. Relevant frameworks may include: • OpenClaw • LangChain agents • OpenAI Swarm • LlamaIndex agents • or similar systems designed for task orchestration The initial agents will support several operational areas. Initial Workflow Areas Project Management Assistant Read information from sources such as: • Outlook email threads • project boards • internal spreadsheets Then: • track job status • identify missing information • highlight upcoming deadlines • generate concise project summaries • produce “next action” lists for project managers The goal is a clear operational snapshot of active projects. Quoting and Proposal Assistance Assist with assembling quotes and proposals using our pricing logic and past projects. Tasks may include: • extracting relevant information from emails or job records • drafting proposal documents or email responses • organizing scope details such as sizes, materials, locations All outputs will initially be reviewed by a human before sending. Billing and Invoice Preparation Help prepare invoice drafts based on completed work or project milestones. Tasks may include: • checking consistency between quotes, purchase orders, and invoices • flagging discrepancies or missing information • generating draft invoices for review The system should surface exceptions rather than silently guessing. Shipping Coordination Generate shipping requests and track shipments. Examples include: • generating labels via UPS / FedEx APIs or similar • preparing shipment information for LTL carriers • tracking shipment status and updating internal records Installer / Subcontractor Coordination Many of our installations are performed by subcontracted sign installers. The agent should assist with: • identifying installers in a given region • requesting installation quotes • tracking responses • collecting insurance documents or required credentials Sign Permit and Code Research For each sign project, we must confirm local regulations. The agent should assist with researching: • city or county sign ordinances • permitted sign types (wall signs, banners, window signs, etc.) • size limits, restrictions, and permit requirements The output should be a clear summary and checklist our team can review. Workflow Matching We already have internal processes for submitting sign drawings and documents to landlords and property managers. The agent should follow these processes by: • ensuring correct file naming conventions • organizing required documents • preparing submissions via email or portal workflows Important: We Value Reliability We are not looking for a flashy demo. We care about: • reliable execution • clear logging and error handling • retry logic when APIs fail • safe handling of business data We prefer structured outputs, validation layers, and deterministic checks over open-ended generation for critical workflows. In other words: production-quality systems rather than experiments. Tech Experience We’re Looking For Strong experience with: • Python or Node.js • building agent-based or automation systems (OpenClaw, LangChain, Swarm, or similar), including tool use, structured outputs, and workflow orchestration • API integrations • automation workflows • email parsing and structured data extraction Experience connecting systems such as: • project management tools (, ClickUp, etc.) • accounting tools (QuickBooks or similar) • shipping APIs (UPS, FedEx, LTL) • email systems (Outlook / Microsoft 365) Helpful but not required: • web scraping for public municipal data • operational automation for small businesses • construction or project-based industries Deliverables For the initial phase we expect: • one or more working AI agents running in our environment • integration with a limited set of workflows • clear documentation explaining: how the system runs configuration and API keys how to extend or add new skills We also ask for a short recorded walkthrough or handoff session. Initial Project Scope We suggest starting with a focused first phase. For example: • project management summaries • installer outreach and coordination • basic billing/shipping support Estimated scope: 20–40 hours over 2–4 weeks If the system works well, we expect to continue expanding it with additional workflows. What We’re Looking For in Your Proposal Please include: Your experience building AI agents or automation systems Links to GitHub projects or real systems you have built Your preferred stack for a project like this Estimated hours and hourly rate for a first phase Location and Collaboration We are based in the United States (Pacific Time). Remote developers are welcome, but you should be able to join occasional calls during U.S. daytime hours. Why This Project Is Interesting Our company runs dozens of projects simultaneously across North America. This means there are many repetitive operational tasks that are ideal for AI assistance. If done well, the system you build will: • run in production • expand over time • become a core part of our internal operations If successful, this system will become a core part of how our company operates day-to-day. We are looking for someone who enjoys building systems that actually get used every day. Additional Screening (Important) To help us find the right fit, please include the following in your proposal. 1. Architecture Question In 4–8 sentences, describe how you would structure an AI agent system to handle: • reading Outlook email threads • summarizing project status from • generating installer outreach emails • preparing invoice drafts Specifically include: • whether you would use one agent or multiple agents • how you would handle tool integrations • how you would prevent hallucinations in financial or client-facing outputs 2. Code Example Please include: • a GitHub repository or • a brief description of a production automation system you have built We are specifically interested in real systems, not demos. 3. Practical Scenario Imagine the system is preparing an invoice draft and detects: • quote = $8,500 • purchase order = $8,000 • internal job record = $8,500 What should the system do? If this type of problem excites you, we’d love to talk.