Shovels

Shovels Inc is now a Sference design partner! Shovels is the intelligence layer for the built world - they turn 150M+ US building permits, contractor records, and government decisions into structured data that property-tech, materials, and trades software runs on. Behind that product sits exactly the kind of workload we built sference for: large-scale LLM classification, normalization, and entity extraction running in production data pipelines. Work that doesn't need a real-time response - it needs to be reliable and cost-efficient at volume. That's what our batch API does. In Petra Kopic's words: "AI has fundamentally changed what's possible with messy real-world data. At Shovels we collect and normalize millions of building permits from thousands of US jurisdictions, and keep pushing the limits of how LLMs turn that data into insight. As our data grows, cost-efficient batch inference on open models becomes key, and we're excited to explore it with Sference as a design partner." — Petra Kopić, VP of Data @ Shovels If your pipelines run classification, enrichment, or extraction at scale and you're paying real-time prices for latency-tolerant work — that's the gap we close. More to come from this partnership.

What was Ryan's first academic love in high school? What opportunity did Ryan uncover by responding to a Craigslist advertisement? Why did Ryan decide to return to his original path after an enjoyable stint in consulting? What attribute of MIT culture appealed to Ryan and convinced him to apply to Sloan for an MBA? What part-time job did Ryan land at MIT that gave him special insight into the campus's eccentricities? How did Ryan end up founding a screen writing software business? How did Ryan pivot this failed idea into a lucrative marketplace business? What hard lesson about unit economics did Ryan learn during this period? How did Ryan connect the dots between data, B2B, and climate policy to start a new tech business? What did Ryan admire about Stripe that he applied to the permit data space with Shovels? Why is collecting and aggregating local building permit data so challenging? What insights has Shovels uncovered around permit activity and different types of climate disasters? What is the permit cloud that Shovel's is uncovering that indicates early signal about a larger infrastructure project? How is Shovels providing insight into available trade resources for intensive projects such as data centers? Ryan Buckley - CEO and co-founder of Shovels, joins Proptech Espresso to answer these questions and discuss how his early childhood interest in roller hockey exposed him to the parking lots of seminal internet titans like Netscape and Silicon Graphics. Listen to our full conversation here: https://lnkd.in/gZ7Rkwjv

AI models require a structured, verifiable grounding layer for spatial retrieval-augmented generation to provide a trusted source of geospatial facts to prevent hallucinations. 🗺️ Overture is the fundamental infrastructure that makes reliable AI innovation possible in the physical world, driving a general industry shift towards collaborative, open, and universally accessible spatial and location grounding for AI and LLMs. In this blog by Albi Wiedersberg, we feature 10 AI startups that are actively leveraging Overture's spatial and location data layers. They are grounding their applications by cross-referencing street-level imagery with Overture's Places and Buildings themes via GERS ID, utilizing GERS to stand up an API turning map listings into interactive AI endpoints, and streaming cloud-native GeoParquet files to power inference-time grounding. Read: https://lnkd.in/eT5CGtgu Zephr.xyz, inHotel, Fused, VOYGR (YC W26), Monarcha, SeerAI, Inc., Atlas.co, Shovels, Aino, Contour