We build data platforms that give operations teams structured intelligence on aircraft anomalies and maritime chokepoint disruptions — before they cascade into costly delays.
Obsidian Dynamics exists to close the gap between when a disruption forms and when the people who need to act on it find out. Whether it's an aircraft holding a suspicious loiter pattern over critical infrastructure or a queue of 400 vessels building at a canal approach 8,000 miles away, our platforms deliver structured, validated intelligence before it becomes breaking news.
We don't sell dashboards. We sell the hours of lead time that let operations teams make decisions instead of react to them.
Obsidian Dynamics is a technical founder-led company. Both platforms — from event-driven data pipeline architecture to the anomaly detection algorithms — are designed and built in-house. No outsourced development teams. No off-the-shelf dashboards rebranded as "intelligence."
The company was incorporated in August 2025 to formalise what began as two independent data engineering projects: one tracking aircraft anomalies using global ADS-B data, the other computing disruption baselines across seven maritime chokepoints from raw AIS position reports.
Both platforms share a common philosophy: ingest live sensor data, compute baselines, detect deviations, and deliver structured intelligence fast enough for operational use. Every architectural decision prioritises speed, auditability, and analyst trust.
Different operational domains, same principle: ingest live sensor data, compute baselines, detect anomalies, and deliver structured intelligence at machine speed.
Browser-first aircraft anomaly monitoring. Eight detection classes — loiter, ghost, squawk, rapid descent, ICAO spoof, GPS jamming, formation flight, callsign duplicate — powered by global ADS-B data with 3D Cesium visualisation.
Learn more about SkyGrid →Maritime chokepoint intelligence at signal speed. Seven critical waterways monitored with AIS-derived delay baselines, anchorage queue tracking, weather-adjusted risk scoring, and route-level risk comparison.
Learn more about Strait Signal →Every detection, alert, and risk score is derived from observable sensor data. No speculation, no editorial. Analysts make the final call — our platforms provide the evidence.
Our detections are advisory signals, not final verdicts. We respect the complexity of real-world operations by flagging anomalies for human review, not making automated judgements.
Everything we build is API-first. Dashboards are important, but the real value is in structured data delivery — webhooks, REST endpoints, and machine-readable payloads that plug into existing workflows.
Intelligence that arrives late is noise. Our systems are designed for 15-minute data cycles and sub-second event routing. Speed is a feature, not a bonus.
Static thresholds break. Our platforms compute rolling baselines from historical data and flag deviations, adapting to seasonal patterns, long-term trends, and evolving operational norms.
Every alert includes context: what triggered it, what the baseline was, and what data sources contributed. Black-box intelligence has no place in critical operations.
We're always interested in speaking with teams who need better intelligence for aviation or maritime operations.