Junior AI Automation Engineer (Claude / Supabase)

Guardian Engineering; Town of Florida (NY); On-site; USD 45,000 - 55,000; Full time

Benefits offered by this job

Profit sharing

On-site work in Jupiter, FL

Job summary

Guardian Engineering is seeking a Junior AI Automation Engineer to join our on-site team in Jupiter, FL, working with Claude Code, Supabase, and GitHub to automate parts of our portal. You’ll read messy real-world documents, transform them into structured data, and wire that into simple tools and scripts.

You’ll contribute to building a scalable data layer, ship customer-visible features, and grow with a small, equity-minded company that rewards ownership as it expands.

Qualifications

• Strong coding in Python and ability to read messy real-world documents.
• Comfort with SQL, APIs/JSON/HTTP for data flow between portals and data stores.
• Some experience using Claude or another LLM for coding or document processing.
• Willing to work in Claude Code as main dev environment; familiarity with Supabase.
• Knowledge of Postgres/Supabase or a strong desire to learn.

Responsibilities

• Own pieces of the closeout process with support: track permits and inspection status across portals and jurisdictions; help schedule/confirm inspections; log results and corrections.
• Assemble closeout packages and chase missing items before delays.
• Build and improve the customer portal: write/refine scripts with Claude Code to pull data from portals and APIs; normalize data into Supabase; use LLMs to extract structured data from PDFs and reports.
• Power real-time status views, alerts, and daily digests in the portal; work in GitHub for version control, PRs, and code review.
• Collaborate with team to ship features that customers actually see and use.

Skills

Python

SQL

APIs/JSON/HTTP

Claude Code/LLM coding

Postgres/Supabase

GitHub

Tools

Claude Code

GitHub

Supabase

PostgreSQL

Job description

Title: Junior AI Automation Engineer (Claude / Supabase)

Location: On-site in Jupiter, FL (atleast 3 days a week)

Who We Are

Guardian Engineering is a licensed Florida private provider firm (F.S. 553.791). We do the plan reviews and inspections building departments do — legally, statewide — except fast. We’ve already built the portal our paying customers use every day. Now we’re automating everything behind it.

Why listen to me: I co-founded a private provider firm in April 2023, built it to 1,000+ projects, and that model recently attracted serious acquisition interest. I know exactly what this becomes when it’s built right. The difference this time: the people building it own a piece of it.

The Role

You’re joining as a junior engineer on the ground floor, working directly in our portal and stack (Claude Code, Supabase, GitHub). You’ll ship things customers actually see and use.

Here’s the target, stated plainly so you can hold us to it: a year from now, this role handles 3x the project volume with the same headcount because of what you built. You will be able to point at that number and say “that was me.”

You don’t need construction experience. You do need to be able to read messy real‑world documents, turn them into structured data, and wire that into simple tools and scripts.

The Deal (no fine-print energy)

• Salary: $45,000–55,000. I won’t pretend that’s a big number. It’s a floor, not the offer.
• The offer: profit sharing that converts to real ownership as the company grows. You’re employee-number-single-digits at a firm run by someone who’s already taken this exact model to an exit-worthy outcome once.
• Budget for AI tooling, API costs, and anything reasonable you need to build with.
• On‑site in Jupiter, FL. If that’s a dealbreaker, this isn’t your job — no hard feelings.
• - If you truly believe you can do this hybrid, apply and follow the requirements below. We are happy to consider, but we need to be wildly impressed by your submittal.

What You’ll Do

Own pieces of the closeout process (with support)

• Track permits and inspection status across portals and jurisdictions
• Help schedule / confirm inspections; log results and correction notices
• Assemble closeout packages and chase missing items before they cause delays

Build and improve our customer portal

• Use Claude Code to write/refine scripts that pull data from portals and APIs
• Normalize everything into Supabase (our main data layer)
• Use LLMs to extract structured data from PDFs and reports
• Power real‑time status views, alerts, and daily digests in the portal
• Work in GitHub for version control, PRs, and code review

What You Need

Technical

• At least one programming language (Python preferred)
• Basic SQL + APIs/JSON/HTTP comfort
• Some experience using Claude or another LLM for coding or document processing
• Willing to live inside Claude Code as your main dev environment
• Familiarity with Postgres/Supabase or a strong desire to learn

Mindset

• Early‑career and hungry for reps on real problems in front of real customers
• You verify AI output, especially on anything that touches regulators
• Bias to ship v1 now and improve, not design for six months

The Deal (no fine-print energy)

• Salary: $45,000–55,000. I won’t pretend that’s a big number. It’s a floor, not the offer.
• The offer: profit sharing that converts to real ownership as the company grows. You’re employee-number-single-digits at a firm run by someone who’s already taken this exact model to an exit-worthy outcome once.
• Budget for AI tooling, API costs, and anything reasonable you need to build with.
• On‑site in Jupiter, FL. If that’s a dealbreaker, this isn’t your job — no hard feelings.
• - If you truly believe you can do this hybrid, apply and follow the requirements below. We are happy to consider, but we need to be wildly impressed by your submittal.

Who should NOT apply

• You want a big-company salary with startup language sprinkled on top
• You need someone to hand you tickets
• You read “profit sharing” and hear “risk” instead of “leverage”

Who should

• You’re early‑career and want reps on real problems in front of real customers
• You’d rather own upside at year one than negotiate PTO at year five
• You verify AI output instead of trusting it — we work next to regulators