The posting, in xAI's own words
archived Oct 5, 2026SpaceXAI’s mission is to create AI systems that can accurately understand the universe and aid humanity in its pursuit of knowledge. Our team is small, highly motivated, and focused on engineering excellence. This organization is for individuals who appreciate challenging themselves and thrive on curiosity. We operate with a flat organizational structure. All employees are expected to be hands-on and to contribute directly to the company’s mission. Leadership is given to those who show initiative and consistently deliver excellence. Work ethic and strong prioritization skills are important. All employees are expected to have strong communication skills. They should be able to concisely and accurately share knowledge with their teammates.
ABOUT THE ROLE:
The Data Center Engineering team designs training and inference compute clusters from first principles, pushing thermal, electrical, and optical systems to the limit of physics. We operate as a highly cross-disciplinary team of mechanical, electrical, optical, and software engineers, all working to maximize the efficiency, reliability, and deployment speed of terrestrial computing to serve rapidly scaling demand. This team owns compute architecture, datacenter network design, rack builds, connectivity, in-house design of our own computers, and the manufacturing line that builds that compute. As an Optical Engineer, you will own the physical optical layer of that computer: fiber layout, link design, connector and polarity architecture, optics selection, and production of a plant that can be built, tested, and repaired at AI cluster scale.
Read the full posting ↓
RESPONSIBILITIES:
Own end-to-end optical and fiber design for AI training and inference fabrics: fiber cabling, cassette/panel architecture, polarity, breakout, reach, and connector strategy (MPO-12/16, SN, MDC, LC, OSFP / QSFP-DD) Write and verify link budgets for scale-out and scale-up interconnect (SMF/MMF, AOC/DAC, DR/FR/LR, InfiniBand and Ethernet), including loss, reflections, FEC, and realistic install margins Develop and implement factory tests for optical interconnects including insertion loss, return loss, and polarization Select and qualify optics, fiber, and connectivity (transceivers, patch, trunks, cassettes, harnesses); own vendor bring-up, failure analysis, and fleet reliability of the optical layer Deploy and verify end-to-end optical interconnect into the manufacturing line Lead commissioning and debug: inspect/clean, power meter, OTDR, BER/FEC stats, and volume failure modes (dirty connectors, wrong polarity, wrong map, damaged end-face) Evolve connectivity toward the next speed and density (800G / 1.6T, CPO/NPO awareness, tighter reaches, higher fiber count per rack) Work with mechanical and network teams to design structured cabling and raceways with the rack and facilities teams: tray fill, bend radius, pre-term vs field-term, labeling, and a plant that can be built and repaired at hall scale
BASIC QUALIFICATIONS:
Bachelor’s degree in electrical engineering, optical engineering, physics, or a related engineering discipline 2+ years of professional experience in fiber layout, optical link design, and/or datacenter/HPC optical engineering
PREFERRED SKILLS AND EXPERIENCE:
Demonstrated ownership of fiber or optical design that was installed and turned up in a datacenter, HPC, or AI-compute environment Working command of single-mode vs multimode, MPO polarity, insertion/return loss, and 100G+ transceiver classes Hands-on optical test: inspect/clean, power meter, OTDR or equivalent Fiber interconnect ownership for AI / high-compute clusters (GPU scale-out fabric, InfiniBand NDR/XDR, RoCE Ethernet 400/800G) High-density pre-terminated trunk/cassette design and rack-to-spine optical mapping at hall scale Optics vendor qualification at volume (OSFP, 800G DR8/2xFR4, AOCs, and the failure modes that show up in a real fleet) Experience designing optical harnesses or cassettes for a manufacturing line, including test coverage and yield DWDM / campus DCI for cluster-to-cluster or building-to-building connectivity Familiarity with CPO/NPO, linear-drive optics, and next-gen packaging as it hits the rack Time at a high-velocity AI cloud, GPU cloud, or custom-compute hardware team