Software Engineer

TAR•San Francisco, CA

17d•Hybrid

About The Position

TAR builds fully islanded solar and storage campuses that power AI data centers, and running a data center around the clock on an island with no grid behind it is a software problem as much as an electrical one. As Software Engineer, you own the software that lets us design, simulate, and operate power campuses faster than anyone else: high-fidelity digital clones of entire campuses, the geospatial data pipelines that feed automated site design, the design automation itself, and the in-house energy management system that will run the island. You will not do all four at once; expect to go deep on one or two and have real influence over how the rest get built. Your work turns into steel and silicon in the desert on a timeline measured in months, with no legacy codebase and no committee in the way. This role is on-site in San Francisco or Austin, with visits to our West Texas site.

Requirements

Strong software engineering fundamentals and a track record of shipping systems that other people depend on.
Fluency in Python and the judgment to know where performance-critical pieces are warranted; language matters less than judgment.
Comfort in at least one of: numerical simulation, geospatial and GIS data, real-time or embedded control systems, or computational geometry and optimization.
The ability to work directly with domain experts, from power systems engineers to civil engineers and project developers, and turn their tacit knowledge into software. You should enjoy learning the physics.
Bias toward first-principles reasoning and toward shipping: we have projects in construction, and software that is not in use is not finished.

Nice To Haves

Prior work in solar, storage, microgrids, or utility-scale power, including familiarity with PVsyst, PSCAD/EMTP, PLEXOS, SAM, or similar.
Experience with SCADA, Modbus/DNP3, IEC 61850, or industrial control stacks.
Experience with CAD or geometry pipelines, or generating engineering deliverables programmatically.
Background in robotics, autonomy, or another domain where software has to survive contact with the physical world.

Responsibilities

Build high-fidelity simulations of an entire campus, spanning PV, inverters, batteries, controls, and the data center load, so sizing decisions, dispatch strategies, degradation assumptions, and outage scenarios get stress-tested before we pour a single foundation.
Work alongside our power systems engineers and outside stability consultants, translating what they know into models that run thousands of times a day instead of once a quarter.
Build data pipelines that ingest and normalize the raw inputs a site design depends on, including LiDAR and DEM topography, geotech and soils data, parcel and constraint layers, irradiance and TMY datasets, and environmental overlays, into a clean geospatial substrate that automated layout can consume.
Turn a site boundary and a set of constraints into a buildable design: grading and racking layouts, DC block topology, cable routing and losses, equipment placement, BOMs, and the drawing artifacts external engineers and EPC partners actually need.
Compress a design cycle that takes months into one that takes days, so we can evaluate hundreds of site variants instead of three.
Help take our energy management system in-house: real-time control, SCADA and telemetry integration, and testing infrastructure for the brain of an islanded campus, which holds frequency and voltage, manages state of charge against a load that never sleeps, coordinates inverters and batteries, and degrades gracefully when something fails.
Own the operational safety work that comes with software controlling physical assets.