About

The team behind Orogen — and the always-on subsurface loop we built it to close.

Orogen is two founders: a seismic-services veteran who has run processing for upstream operators for two decades, and an applied-ML engineer who reads seismic and well telemetry for a living. Both have watched real prospects slip because the next quarter’s batch window was too slow. Always-on subsurface agents for upstream oil and gas — ingest seismic and well telemetry, flag anomalies, and auto-write prospect briefs, reservoir reports, and drilling-risk alerts.

Talk to us Walk the pipeline

Mission

Subsurface intelligence can't wait for the next batch window.

The signals that decide an asset — a 4D seismic repeat, a pressure-rate transient, a completion signature in the WITSML stream — arrive continuously. A quarterly reprocessing cycle, or even a weekly batch, loses those signals by definition. The prospect that mattered was the prospect that showed up between cycles.

Orogen closes that loop. Continuous seismic ingestion through SLB TruLink and Stream, live telemetry over WITSML, and a fleet of small agents doing cross-correlation, AVAZ binning, decline-curve fitting, and transient convolution in the background. The output is a brief written for the squad that owns the call — not a dashboard that waits to be read. Detection to briefed queue, in seconds instead of weeks.

Founders

Two founders. One has run processing. One has built the model.

The split is deliberate: the operations bar is set by someone who has signed off on crew handoffs in the field, not by a model vendor demo dataset. The engineering bar is set by someone who has shipped cross-correlation and transient convolution against real instrument-grade data.

CEO

Founder, Operations

Two decades running seismic processing for upstream operators — time-lapse (4D) re-processing across the North Sea and Gulf of Mexico, transition zones where the field data is genuinely messy, and the long-tail patches that never make it into a sales deck. Has signed off on more crew handoffs than they can count, which is the reason Orogen treats the inbound stream as the source of truth instead of a polished SEG-Y re-export.

Background: Seismic services · processing lead
Field reality: 4D + transition-zone reprocessing
Operating instinct: Crew handoffs, vendor SLAs, audit trails

CTO

Founder, Engineering

Applied ML on instrument-grade subsurface signals — 3D and 4D seismic cross-correlation, AVAZ binning against well-test priors, decline-curve fitting on noisy commingled production, pressure-rate convolution over WITSML transients. Built the in-house pipeline behind the agent fleet: the parts that distinguish a synthetic demo set from a streamer file with a blown near-offset channel.

Background: Applied ML · applied geophysics
Phenomena: AVAZ, decline curves, transients
Stack: Cross-correlation · binning · time-index

How we got here

Built for the messy stream — not the polished demo set.

Orogenstarted inside a processing team that kept losing prospects to its own scheduling. The synthetic data the model vendors shipped was always clean — and the real stream almost never is. Blown near-offset channels, missing SIBR frames, commingled production with no clean split. The first version of the agent fleet was written against that operator's data, not a benchmark, and it's been hardened against field realities ever since: instrument-grade inputs, audit trails on every detection, and a brief the squad that owns the call can actually act on.

Talk to the team

Send a well, a survey, and a friction point — get a briefed queue back the same week.

Direct line

Email orogen@polsia.app

Or read how the pipeline runs first.