Andreessen Horowitz has recently invested in a stealth startup building a space-based data transport network that moves enterprise data between two points on Earth using free-space laser links in MEO. The company is focused on providing multi-Tbps capacity network infrastructure, relying on optics and unique constellation orbit design to deliver reliable bandwidth to enterprises and sovereigns. If you are interested in learning more, please apply or contact the a16z team to connect with the founders.
You aim and lock the beam. For the a LEO demonstrator you acquire a satellite and hold the beam on it across a pass measured in minutes, not hours — high angular slew rates, significant Doppler, and fast, reliable acquisition are the game. Nothing transmits until you've closed and held the loop.
The MEO end state is long, turbulence-dominated passes at high MEO, where the beam is narrow enough that small errors result in us missing the target entirely. We want someone who can close the LEO link now and architect the loop to stretch to MEO. You own the optical and fine-steering side of the loop and beam lock. The GNC – Flight Dynamics & Navigation engineer owns platform pointing, orbit/ephemeris, and acquisition geometry and hands off to you at acquisition; the GNC – Simulation & Analysis engineer gives you the 6-DOF and end-to-end pointing environment you design against; a later Controls Engineer deepens fine steering and fiber coupling for production.
The closed-loop acquisition, tracking, and fine-pointing architecture for the demonstrator, built to stretch to MEO.
The acquisition sequence and timeline against a fast-moving target.
Fine-steering-mirror control and beam lock through atmospheric turbulence.
Doppler and high-slew-rate tracking strategy for short LEO passes.
The optical-side pointing error budget and the acquisition handoff with GNC.
The optical pointing loop's path from LEO-demo passes to MEO-class passes, in coordination with the modem EE on real-time control implementation.
Closed-loop pointing and tracking control depth — your home turf.
Fast/fine-steering-mirror or line-of-sight stabilization control experience.
Acquisition strategy for moving targets.
A feel for sensor/actuator latency, loop bandwidth, and disturbance rejection.
Strong modeling skills (Python, MATLAB, or similar) and hands-on lab experience.
Lasercom / free-space optics pointing, or LEO / moving-target tracking with Doppler.
Adaptive optics, wavefront sensing, tip/tilt sensing, or beam-stabilization hardware.
Directed-energy or electro-optical target-tracking / gimbal line-of-sight stabilization background.
Real-time control implementation on FPGA or embedded targets.
If this list doesn't perfectly match your background but you've done serious closed-loop pointing-and-tracking work, reach out anyway — we hire for the skill, not the acronym, and we are always looking for exceptional people.
The company is a lean company by design. There is no program office, no requirements team handing you specs, and no army of analysts. You will build the models, make the calls on incomplete data, and be forced to defend and revise them. You will spend meaningful hands-on time in the lab, not just in reviews. You will own your piece of the system and live with the consequences of your own trades. If you do your best work inside a large, well-resourced program, this seat will frustrate you. If you have been waiting for that kind of ownership from the start, this is the job for you.
The company offers competitive, best-in-class compensation including salary, meaningful equity, and full benefits. Health, dental, vision, generous paid time off (PTO), relocation support, and a real equipment and lab budget.
Location: Culver City, CA (onsite).
