Senior Geospatial Perception Engineer

About Every core aspect of society, from our seaports to our airports to emergency services depend on 31 GPS satellites. The US economy will lose over 1 billion dollars a day if GPS is disrupted. In January 2022, the entire Denver airport had its operations jammed for 33.5 hours, impacting the aircraft and control towers, and everything in between, all due to a small jammer a man had in his work truck; furthermore, the Russia-Ukraine war has escalated GPS jamming incidents. This is why we are building the alternative to GPS: an offline solution that enables any person or robot to locate themselves on Earth offline at a high level of accuracy (≤ 6 inches) using LiDAR fingerprints. Over the past 4 years, our mission has evolved to interpret the natural characteristics of the earth (including the built environment) to enable precise geolocation without GPS or wifi. Our Story As refugees from Vietnam, my family traveled around the states in search of work, from Washington, Maryland, Georgia, and all the way to St. Louis. Navigating new places, communities, and cultures was difficult. Furthermore, GPS technology completely failed when it was absolutely critical: subways, airports, and downtown corridors. In 2021, I took this hardship as inspiration and turned it into the founding of Xphera, a personalized tool and visual GPS (now xEarth) to help displaced individuals and refugees navigate unfamiliar spaces with confidence. From our work in the navigation space, we discovered the multi-trillion dollar risks of GPS and the pervasive dependency of our entire global infrastructure. Learn more: https://www.xphera.earth/ Position Overview XPHERA is building a DEM-based alternative to GPS. Our navigation engine ingests LiDAR and other range sensors, matches them against high-resolution DEM, LiDAR, and multibeam bathymetric (sonar) maps, and outputs WGS84 positions via standard NMEA sentences. OEMs can integrate XPHERA as a drop-in GPS replacement while retaining full control over vehicle reference frames and lever arms. As Senior Geospatial Perception Engineer, you will own the core algorithms that turn range sensor data into a precise navigation solution. You will design the “terrain fingerprints” and map-matching logic that allow our engine to estimate the position and velocity of the XPHERA sensor frame in WGS84, and you will work closely with the team to deliver this as a robust GPS-alternative SDK that outputs NMEA to OEM customers. Key Responsibilities • Design and implement DEM/LAZ “fingerprinting” pipelines (tiling, feature extraction, descriptors) for large-scale map databases. • Build LiDAR and range-sensor map-matching algorithms that align live point clouds with DEM/LAZ/sonar fingerprints to estimate pose in Earth or map coordinates. • Work on the navigation core that fuses map-matching constraints with IMU (and optional aiding) to produce a stable, drift-bounded WGS84 solution suitable for NMEA output. • Define and manage reference frames: sensor frames, vehicle body frame, map coordinates (projected CRS), and conversion to WGS84/ECEF for external interfaces. • Develop and maintain large-scale indexing and retrieval for map fingerprints (for example, using approximate nearest neighbor search) under strict latency and memory constraints. • Prototype and evaluate optional monocular image/depth fusion paths that can assist or back up LiDAR-based localization when needed. • Own metrics and benchmarking for navigation quality (position error, integrity, convergence time, degradation behaviors) and drive experiments to improve performance. • Collaborate with SDK and systems engineers to expose a clean C-style API and NMEA output that OEMs can integrate with minimal changes to existing navigation stacks. Qualifications • Strong background in computer vision, robotics perception, or navigation with direct experience in at least one of: SLAM, visual/LiDAR localization, map-matching, or HD map-based navigation. • Hands-on experience with geospatial data, including DEM/DSM rasters and/or LiDAR point clouds (LAS/LAZ), and familiarity with common coordinate reference systems. • Solid understanding of 3D geometry and pose estimation (for example, coordinate frames, ECEF/ENU, PnP or ICP, state estimation concepts). • Proficiency in Python and C++ and comfort working with performance-sensitive code and large datasets. • Familiarity with tools such as GDAL, rasterio, PDAL, or PCL, and with at least one deep learning framework (for example, PyTorch or TensorFlow) for learned descriptors if applicable. • Experience designing or working with feature descriptors, retrieval systems, or approximate nearest neighbour search for large-scale matching problems. • Ability to reason clearly about reference frames, lever arms, and the distinction between sensor frames and vehicle reference points in OEM integrations. Nice to Have • Experience with GPS/INS integration, GNSS receivers, or other satellite navigation systems, especially where map-matching or augmentation is involved. • Background in geodesy or high-precision mapping (for example, WGS84, ECEF, UTM, state plane, datum and height model basics). • Experience with embedded or real-time systems in vehicles, drones, or robotics platforms. • Prior work on terrain analysis, topographic feature extraction, or LiDAR-based surveying and digital twins. • Experience working with sonar / bathymetric data, (i.e. NOAA NCEI’s Multibeam Bathymetry Database (MBBDB) or similar datasets), and designing fingerprinting or matching pipelines over bathymetric DEMs that leverage attributes like backscatter intensity and multiple returns/echoes as part of the descriptor. How this role is different This is not a generic computer vision role. You will be working at the intersection of geospatial mapping, robotics perception, and navigation systems, designing a terrain-based GPS alternative that OEMs can integrate through a familiar NMEA and SDK interface. Your work will directly shape how XPHERA defines and delivers a new class of map-first positioning to vehicles, aircraft, and infrastructure operators.