AI Automation Specialist

Job Description

Job Description

Overview

Upward Mobility with Propark Mobility!

 

AI Automation Specialist

Propark Mobility Corporate Headquarters

 

$70,000 - $80,000 Annual Salary Range

Full-time Opportunity; Full Company Benefits

 

Work Modality: Hybrid

Schedule: Full-time, Monday - Friday

Location: Hartford, CT

 

Snapshot  

• The company: Propark Mobility is a national parking and mobility company, founded in Hartford in 1984 and now operating over 1,000 locations in more than 250 cities. 
• Role: One of the first hires in a new department. You’ll build and ship, choose the stack, and set the standards the next hires build on. 
• What you do: learn a manual business process from the person who runs it, document it, then replace it with AI tools and custom agents. 
• The bigger arc: help unify Propark’s data and build AI into its own software, alongside the AVP of AI. 
• The mix it takes: enough technical skill to build custom agents, enough business sense to understand how a manual process actually works, and enough polish to train non-technical staff on what you ship. 

Best fit: someone who came up through a business or analytical background and then taught themselves to ship real AI systems. 

About the work  

The AI Department turns Propark’s manual, expert-run back-office work into software and agents. This is the kind of work that today lives in one person’s head and a stack of spreadsheets.

For example: 

• Reconciling two systems that are supposed to agree, and flagging every mismatch. 
• Pulling data from several sources into a report that someone rebuilds by hand every month. 
• Checking incoming documents against records before anything gets approved or paid. 
• Turning a recurring spreadsheet routine into a tool that runs on its own. 
• Pulling structured data out of PDFs, emails, and forms so no one has to retype it. 

Work that eats a full day every month should take minutes. Closing that gap, across one process after another, is where you start. The bigger arc, which you grow into alongside the AVP of AI, is unifying Propark’s data and building AI straight into the company’s own software. 

What you’ll do  

• Sit with the expert who owns a manual process and learn their workflow the way you’d reverse-engineer an undocumented codebase. Ask the obvious questions, watch them work, and find the edge cases they handle without thinking. 
• Document the process: inputs, steps, decision rules, and exceptions. If you can’t write it down, you can’t automate it. 
• Build the automation with AI coding tools as your daily drivers: Claude Code, Cursor, and Codex. 
• Design and ship custom agents for specific jobs: retrieval over internal documents, agents that chain tools and make decisions, and tools that replace hours of manual review. You own them from first idea to production. 
• Ship to real users inside the company, watch the work break, and fix it until the process runs without you. 
• Run AI training sessions for non-technical colleagues so the people who use what you built can run it without you. 
• Help set the department’s direction: the stack, the standards, the roadmap, and what we take on next. 

Bigger projects  

Process automation is the on-ramp. As you find your footing, you’ll take on larger work with the AVP of AI: 

• Break down data silos and help stand up a central data warehouse, so every automation and agent works from one source of truth instead of scattered exports and one-off reports. 
• Build AI into Propark’s existing proprietary applications, not just standalone tools alongside them. 
• Prototype and push new ideas with the AVP of AI, testing what becomes possible as the department and the tools mature. 

How we work  

• Business first. We don’t automate a process until we understand it. 
• AI coding tools are the default here. 
• Success is measured in processes shipped and hours handed back to the business. 
• Small team, fast decisions, direct access to the people running it. 

What we’re looking for  

• 1 to 4 years building real software, including hands-on AI/ML work. 
• Shipped at least one of these yourself: a retrieval (RAG) system, an LLM-powered tool, or a working agent. 
• Real business sense: you can sit with someone who runs a manual process, follow what they describe, and see how to turn it into something a machine can run. 
• Full-stack fundamentals: Comfortable using AI to code in a modern web stack, enough to take a tool from idea to deployed on your own. 
• Working knowledge of a modern AI stack: an orchestration approach, a vector store, and a major model provider. We care that you’ve built the real thing, not which library you used. 
• AI coding tools already in your daily workflow (Claude Code, Cursor, Codex, Github Copilot, or similar). 
• Strong communication. You can stand in front of a non-technical room and explain AI clearly and patiently, without jargon. That matters here as much as the code. 
• A bias for finding the work instead of waiting for it. 
• US-based, able to work a hybrid schedule in Hartford, CT, with regular days on-site. 

Nice to have  

• A business or analytical background, or hands-on experience with manual back-office processes such as reconciliations, recurring reporting, or compliance. 
• A portfolio of self-directed builds, side projects, or deployed apps. 
• Data engineering experience: SQL, building data pipelines, or standing up a data warehouse (for example, Snowflake, BigQuery, or Redshift). 
• Cloud deployment with Azure, GCP, or AWS (containers, managed databases, serverless). 
• Classic ML and data science (Pandas, scikit-learn, XGBoost, explainability tools like SHAP or LIME) for automations that go past simple model calls. 
• Experience presenting to or training stakeholders. 
• Enterprise or regulated-environment experience, where careful data handling and internal tooling at scale matter. 

 

What We Offer

• Fa