Software Engineer - Systems

Design and build low-latency networking infrastructure connecting embedded devices and cloud systems — protocol design, congestion handling, and tuning for throughput and reliability across a distributed sensor network

Build resource-efficient pipelines to ingest and egress multimodal sensor data and telemetry, handling packetization, buffering, and backpressure across constrained device environments

Own low-latency command and control infrastructure across a distributed sensor network, with a focus on fault tolerance, deterministic timing, and graceful degradation

Integrate and fuse multimodal data streams from cameras, IMUs, and other sensors — working across driver boundaries, synchronization, and calibration to produce reliable inputs for downstream algorithms

Build and optimize video and image processing pipelines end-to-end: capture, hardware-accelerated encode/decode, streaming, and storage

Contribute to tracking and state estimation algorithms, bridging raw sensor data and meaningful system outputs in close collaboration with ML and perception teams

Build and maintain CI pipelines, test harnesses, and reliability tooling — the simulators and replay systems that let the team move fast without breaking things in the field

Instrument, profile and benchmark system performance — CPU/GPU utilization, memory pressure, network throughput and latency — and drive systematic improvements

Broad systems experience across the areas below, with demonstrable depth in at least one — whether that's networking, video/sensor pipelines, or low-level Linux systems work

Production Rust (preferred) or C++ in low-latency, embedded, or systems contexts — with real ownership of performance, reliability, and resource constraints

Deep networking knowledge (UDP, TCP, QUIC) beyond the API level — packet loss, flow control, retransmission, and tuning for real-world conditions; strong Linux systems fundamentals including IPC, scheduling, and memory management

Hands-on hardware integration experience — cameras, IMUs, or other sensors — including driver interfaces, kernel boundaries, and video pipelines (capture, encode/decode, streaming via V4L2, GStreamer, FFmpeg, or similar)

Proficiency with concurrency and parallel programming — lock-free structures, async runtimes, thread management — with a track record of shipping correct, performant, concurrent code

Comfortable owning CI infrastructure, test harnesses, benchmarking pipelines, and observability tooling alongside feature work

Experience working alongside real-time, multimodal ML data ingestion systems — understanding the data quality, latency, and throughput requirements that make or break model performance

Hands-on experience with modern video codec implementations (H.264, H.265, AV1) across hardware platforms — encoder tuning, rate control, and platform-specific acceleration (V4L2, NVENC, etc.)

Robotics, perception, or state estimation background — familiarity with sensor fusion, localization, tracking algorithms

Experience writing Rust and Nix