Overview

Mercator Technologies (YC S22) was a San Francisco-based startup founded in 2022 by Andrew Duberstein and Dayton Thorpe, two data practitioners who met at Instacart.The company attempted two distinct product strategies in under two years: first, a geospatial fleet management suite targeting vehicle operators, and then a full pivot to Dubo, an AI-powered SQL editor.

Despite genuine technical depth—including a peer-reviewed arXiv paper and a claimed state-of-the-art result on the BIRD-SQL benchmark—Mercator never documented meaningful commercial traction, never grew beyond two employees, and raised only approximately $500K in confirmed funding.The company wound down quietly sometime after April 2024, with no public shutdown announcement.

Its core failure was structural: a two-person team with thin capital attempted to compete first in enterprise geospatial SaaS, then in AI developer tooling, both markets where well-funded incumbents and rapidly commoditizing foundation models made durable product-market fit nearly impossible to achieve.

Founding Story

Mercator's founding team brought credentials that were, by any measure, unusually strong for an early-stage startup.

Andrew Duberstein spent six years at Uber, accumulating experience across three distinct roles. He began as a data scientist detecting fraud in Uber's China business, then moved into engineering management on the self-driving cars program, and ultimately built pydeck—a geospatial data visualization library that went on to accumulate over 12 million downloads. ^{[1] [2]} That combination of applied ML, autonomous vehicle engineering, and open-source geospatial tooling gave Duberstein a rare vantage point on the intersection of location data and developer infrastructure.

Dayton Thorpe arrived from a different direction. He earned a Physics PhD at UC Berkeley, then moved through a sequence of data science roles: Kabbage (2017–2018), Valor Equity Partners (2018–2020), and Instacart (2020–2022). ^[3] He also holds a patent for optimizing insulin dosing for type 1 diabetics with automated insulin pumps—an indicator of applied quantitative range well beyond typical data science backgrounds. ^[4]

The two met on the Instacart data science team, where the shared experience of working with messy location and logistics data appears to have planted the seed for Mercator. ^[5] The company was accepted into Y Combinator's Summer 2022 batch, though it entered the program under a prior name—"Quincy's"—suggesting the founding thesis was still being refined even as the batch began. ^[6]

At the conclusion of the S22 batch, Duberstein announced the company publicly on LinkedIn: "Excited to announce Mercator with my co-founder Dayton Thorpe as we finish up the Y Combinator S22 batch." ^[7] The initial product was described as a suite of mapping tools for vehicle fleets—real-time tracking, custom geofencing, and a GPS reliability developer API. The framing drew directly on Duberstein's pydeck experience and Thorpe's logistics data background.

What the founding story does not explain is why the company was called "Quincy's" before rebranding, what that earlier concept entailed, or precisely when the decision was made to build for fleet operators specifically. The pivot from "Quincy's" to "Mercator" to "Dubo" traces a company that was searching for its identity from the very beginning.

Timeline

• 2022 — Mercator Technologies incorporated in Delaware, headquartered in San Francisco. Company was previously operating under the name "Quincy's" before rebranding. ^[6]

• June 2022 — Mercator accepted into Y Combinator's Summer 2022 (S22) batch. Founders Dayton Thorpe and Andrew Duberstein, who met on the Instacart data science team, begin the program. ^[5]

• August 2022 — Mercator raises a Pre-Seed round led by Y Combinator (~$500K per Tracxn). Investors include Tribe Capital, Soma Capital, Valor Equity Partners, Amplify Partners, and Rebel Fund. ^{[8] [9]}

• September 2022 — Andrew Duberstein announces Mercator publicly at the conclusion of the YC S22 batch. Initial product described as a suite of mapping tools for vehicle fleets: real-time tracking, custom geofencing, and a GPS reliability developer API. ^[7]