Staff SWE - Networking

Somana is a generative robotics company. We founded the company because the incredible advances in the modeling layer have no way to access a fragmented physical world. We're building the connective tissue that brings physical AI into the industrial world.

Somana is a hardware, software, and services platform powering rapid improvements in capacity, throughput, yield, and quality. Our compute platform will talk to hundreds of peripherals, survive factory floors, and keep getting smarter - which means the network fabric that binds modems, WiFi, sensors, cameras, switches, and compute subsystems together is as critical as the silicon itself. If you have a passion for solving hard distributed systems problems under real-world constraints — bandwidth-limited RF links, dynamic topologies, strict latency requirements — then this role is for you.

ROLE OVERVIEW

As a Staff Networking Engineer, you will own the software architecture and hands-on implementation of the core networking and communication systems that power our family of compute products. You will be the subject-matter expert for all things L1-L4, who still reads schematics, reviews driver code, debugs packet traces, and cares deeply about interrupt latency. You will be the one who knows whether a timing issue is in the PHY, the driver, or the scheduler - because you've been in that exact fight before on products at scale.

CORE RESPONSIBILITIES

Design and improve the networking software stack, from prototype through high-volume production and field deployment, to operate reliably, securely, and compliantly

Define and own the L1-L4 networking architecture across heterogeneous interfaces

Manage scalability and performance tuning of multi-interface networks under real-world factory conditions; optimize for bandwidth, latency, jitter, and power

Serve as the primary technical interface between software, electrical, and firmware engineering teams, ensuring that networking constraints are resolved as a unified system across varied environments

Drive hardware bring-up and validation of networking subsystems, from BSP integration through driver development, protocol tuning, and production diagnostics

Own field reliability of the network stack: packet loss budgets, failover logic, HA, OTA pipelines, and degraded-mode behavior when factory conditions are not perfect

Establish design standards, documentation, and code review norms appropriate for a team that is moving fast but shipping hardware that has to last

PRIMARY QUALIFICATIONS

Bachelor's or Master's degree in Computer Science, Electrical Engineering, or related

Proficiency in a systems programming language (Rust, C, C++, or Go)

5+ years of professional experience designing, implementing, and shipping networking, robotics middleware, or embedded software systems at scale

Demonstrated track record of owning network architecture across multiple physical interfaces - you have shipped products that had to coordinate cellular, WiFi, Ethernet, and/or sensor/camera links under one software stack

Deep hands-on experience with Nix or Linux networking internals, user space development, and multi-threading or concurrent programming

Proven ability to debug at the hardware/software boundary: protocol analyzers, oscilloscopes, JTAG, packet traces, and vendor BSPs are tools you have actually used to solve real problems

Strong systems intuition and communication skills

PREFERRED SKILLS AND EXPERIENCE

Experience building networking systems for compute or edge platforms that operate in challenging physical environments: industrial IoT, energy storage, wireless mesh networks, or cloud-edge connectivity

Comfort with manufacturing and production processes for networking hardware: RF validation, certification awareness (FCC/CE), DFM considerations for antenna and connector placement, and production test harness design

Enjoy reading / research in the field of robotics, computer vision, sensor fusion, or similar

LOCATION & COMPENSATION

San Francisco, CA OR SoHo, New York City