Data Quality & AI Readiness Intern

About ArtHaus Properties

ArtHaus Properties is the property management platform built inside ArtHaus, a Bay Area real estate firm with a 45-year operating track record. We manage 2,500 residential units across 50 buildings in Oakland and Berkeley, scaling to 4,000 units over the next 24 months. Our vision is to redefine property management as a resident-obsessed, tech-enabled craft.

About this role

The single biggest unlock for our operating model over the next three years is a property portfolio whose data is clean, complete, and AI-ready. Yardi, HappyCo, EliseAI, and several adjacent systems hold the information we need to deploy AI agents across leasing, maintenance, collections, and asset management — but the data is uneven. Some properties are pristine. Others have gaps. Equipment, units, tenants, vendors, leases — each entity has its own shape of mess.

This internship exists to fix that, end-to-end. You will decide what “complete” looks like, walk the properties to find what’s missing, get it into our systems with the right taxonomy, and then prove the thesis by running real AI experiments on the cleaned dataset.

What you’ll work on

Phase 1 — Data audit (weeks 1–3)

• Map what data we have across Yardi, HappyCo, and adjacent systems for our 36 properties
• Define the “complete record” for each entity: building, unit, tenant, equipment, vendor, lease, work order
• Build a portfolio-wide gap map showing where each property stands and what is missing

Phase 2 — Field collection and system entry (weeks 4–8)

• Visit properties across our Oakland and Berkeley portfolio
• Inventory equipment, verify unit specs, confirm vendor relationships, fill missing records alongside our RPM and maintenance teams
• Enter data into Yardi and HappyCo using a consistent, queryable taxonomy

Phase 3 — AI pilots (weeks 9–12)

• Select one or two high-impact use cases with the leadership team (candidates include maintenance prediction, lease anomaly detection, vendor performance ranking, or turn-time forecasting)
• Build a working pilot against the cleaned dataset
• Present findings and a productionization roadmap directly to the President and operating leaders

What we’re looking for

This role is open to a broad range of backgrounds — we screen on portfolio, skills, and how you think, not your major or program. You’re a strong candidate if you have:

• Comfort wrangling messy data in Python, SQL, or a serious spreadsheet workflow
• Genuine curiosity about how AI actually gets deployed in operating businesses, not just papers and demos
• Willingness to do the unglamorous work — walking properties, talking to maintenance techs, reconciling records — that makes the glamorous work possible
• A bias toward finishing things, not just analyzing them
• Bonus points for any of: prior exposure to property management or PropTech, hands-on experience with LLM APIs, or a portfolio piece where you cleaned data and shipped something real.

What you’ll get out of it

• Direct project ownership with weekly access to the President and senior operators
• A real operating problem at a real operating company — your work goes into production, not into a deck that dies after demo day
• A portfolio-grade case study (cleaning the data layer for an AI-first property management company is genuinely rare experience)
• Exposure to the full real estate operating stack — leasing, maintenance, finance, asset management — at one of the Bay Area’s most active multifamily operators

Logistics

• 10–12 weeks, full-time, Summer 2026
• Oakland-based, hybrid (most field days in-person, deep-work days flexible)
• Reliable car and valid driver’s license required (mileage reimbursed)
• Paid hourly — compensation commensurate with experience