Founding Team, Data Operations & Analyst

FilmoreAI is the intelligence layer for the construction equipment industry a multi-hundred-billion-dollar economy where dealers manage every stage of the machine lifecycle (acquisition, financing, utilization, service, trade-in, disposition) on data that lives in a dozen disconnected systems and a thousand reps' heads.

We're building the data and AI system that fixes that. We’re building equipment domain specific reasoning using a propietary ontology that connects ERP work orders, CRM opportunities, OEM telematics, UCC filings, auction results, and DMS transactions into a single canonical model of every machine, every customer, every dealer interaction across the lifecycle. Aftermarket, where dealers earn the majority of their profit on tribal knowledge is where the data is messiest and the leverage is highest, so it's where we lead.

The data system is the product. We parse public information across all 50 states including UCC liens, construction projects, contractors, and early land development. We normalize telematics across OEM standards. We resolve entities across systems that have never spoken to each other.

Pipeline operations & data quality. Run the daily and weekly heartbeat of the data system. Monitor 50+ scrapers across UCC vendors, county registries, permit systems, telematics feeds, and enrichment APIs. Triage failures fast rate limit vs. schema drift vs. expired credential vs. real bug and route accordingly. Maintain quality gates and dashboards. Catch silent failures (status=‘success’ with 0 rows loaded) before an engineer or a dealer does.

Data acquisition & county outreach. Expand coverage by getting data we can’t scrape. Call county clerks and recorders to negotiate UCC filing access FTPs, bulk exports, secretary-of-state portals. Manage vendor relationships (Accutrend, Info Evolution, Shovels) and the credential and renewal calendar. Document every source what, where, who, cost, refresh cadence so the team trusts the inventory without asking. You are the human interface to the parts of the system that aren’t APIs yet.

Analysis & dealer-facing signals. Translate raw data into answers. Run ad-hoc SQL against Supabase Postgres to answer the questions reps and the founder ask: which counties spiked this week, which contractors are expanding, which UCC filings landed against which dealers’ customers, which leads scored above threshold. Run the lead-quality QA loop so the platform doesn’t ship junk to reps. Pull samples, eyeball them, and write the one-paragraph explanation a sales manager will actually read.

Data & Pipelines

Python, Playwright, public APIs; Airbyte + custom Python connectors; BigQuery (warehouse) Cloud SQL Postgres with pgvector

Reasoning & Agents

Model Router across Claude, GPT, Gemini; LangGraph for agent workflows; internal MCP tool registry; Postgres action ledger

Cloud & Backend

GCP Cloud Run, Temporal Cloud (orchestration + human-in-the-loop), Python + FastAPI, Terraform, GitHub Actions, Secret Manager

Delivery & Observability

Twilio (SMS), native CRM APIs, Retool on BigQuery; Datadog, LangSmith

• SQL fluency (3+ yrs). Joins, window functions, CTEs without Googling. You read EXPLAIN well enough to know why a query is slow. Postgres preferred; BigQuery a plus.

• Python (2+ yrs). Not full systems engineering but you can read a scraper, fix a broken selector, parse a weird CSV, write a one-off backfill. Comfortable with requests, pandas, and a messy notebook.

• Excel and spreadsheets at a senior level. Pivots, lookups, conditional logic, basic modeling. Half this job lives in files counties and vendors send you.

• Operational discipline. You run a weekly checklist without drift. You log what you ran, what failed, what’s pending follow-up. Your handoff notes don’t generate three questions.

• Outreach and negotiation. You can call a county clerk in Florida, explain what we need in plain language, and get to a yes or a routed contact. You ask a vendor for a sample file without burning the relationship if they say no.

• Strong plus: government and public records experience (UCC, secretary of state, permits, court records), data QA and lead scoring, dbt basics, geospatial fundamentals (lat/lng, FIPS).

• You operate without supervision. We don’t ticket every county to call. You see the gap, take the call, write the recap.

• You sweat the small stuff. A scraper returning 0 rows is not passing. A duplicate filing is not loaded. You catch the off-by-one before it ships to a dealer.

• You navigate ambiguity. A county tells you no, a vendor file is half-broken, the schema you queried last week changed. You adapt and route around it instead of escalating after one try.

• You bias toward action and toward asking. When you hit a wall on schema or auth, you escalate fast — you don’t sit on it for two days. When you find something weird in the data, you flag it before it propagates.

• You care about why this exists. Dealers run on tribal knowledge. Every county you bring online and every bad load you catch directly compounds the moat. If that mission doesn’t pull you forward, the rest of this won’t.

AI-Native

• You use Claude or Cursor as your daily driver. SQL drafting, regex, error log triage, outreach email drafting, navigating a county website that hasn’t been updated since 2008 Claude is always open in another tab. You move 3-5x faster than a peer who isn’t using it, and you know it.

• You design context, not one-shot questions. When you ask Claude to write a query, you paste the schema, an example of good output, and the failure case you’re guarding against. You build reusable prompts for recurring tasks county outreach templates, scraper failure triage, weekly QA reports.

• You let agents do the boring work. Backfill scripts, log parsing, data sampling, draft emails, meeting recaps you delegate to the model and review like an editor. You don’t hand-write what the model can draft in 30 seconds.

• You make your work reusable. When a prompt works, you save it. When a task repeats, you templatize it. The next person on the team or future you, three weeks from now picks it up without re-deriving.

• You know when not to trust the model. Numbers in a dealer-facing report get verified manually. SQL on production data gets read line-by-line before run. Outreach emails get a human pass before send. Calibrated, not credulous.

• Small team, direct founder access, decisions get made fast. Async-first clear written updates and run summaries, no standup theater.

• Competitive market comp for a senior data analyst and operations profile. Open to contracting and scope or hours if preferred.

• Preferred location in Texas (Austin) or Lousiana (New Orleans)