> Markdown version of [/jobs/ext/2953734-staff-software-engineer-neuronavigation-and-autonomy](https://www.wearedevelopers.com/jobs/ext/2953734-staff-software-engineer-neuronavigation-and-autonomy). Every page supports `.md` or `Accept: text/markdown`. Links point to the HTML versions so they work for humans too. Agent guide: [/agents.md](https://www.wearedevelopers.com/agents.md). --- # Staff Software Engineer, Neuronavigation and Autonomy - **Company:** Magnus Medical Group LLC - **Location:** Burlingame, CA, United States - **Experience:** Expert - **Salary:** $170,000.0 - $220,000.0 - **Contract:** Permanent contract - **Skills:** Application Programming Interfaces (APIs), Data Analysis, Computer Vision, C++ (Programming Language), Nvidia CUDA, Linux, Image Registration, Python (Programming Language), Kinematics, Machine Learning, Sensor Fusion, Software Engineering, Data Streaming, Strategies of Testing, Multithreading, Real Time Systems, State Machines, Low Latency, ONNX (Open Neural Network Exchange) Format, TensorRT, C++14 - **Published:** September 17, 2026 - **Apply:** https://startup.jobs/staff-software-engineer-neuronavigation-and-autonomy-magnus-medical-10106120 ## About the Role * MSc. degree in computer science, engineering, physics, applied mathematics, or a related technical field, or equivalent practical experience * Expert-level modern C++ with 6+ years delivering production real-time systems: multi-threaded and concurrent design, deterministic low-latency execution, memory and cache-aware optimization, and profiling under hard timing constraints * Proficiency in Python for algorithm prototyping, data analysis, and tooling * Depth in at least one of the following, with the appetite to build competence in the others: geometric computer vision and 3D perception; numerical methods for spatial estimation and sensor fusion; robotics, including kinematics and real-time control; machine learning for perception * Working knowledge of camera calibration, registration, multi-view geometry, rigid-body transforms, and coordinate-frame reasoning * Experience building deterministic data flow over asynchronous sensor streams on Linux, including buffering strategies and time synchronization across devices * Experience with closed-loop or autonomous systems, including state machines, degraded modes, and safety-aware design * A measurement-driven approach to accuracy and robustness, including designing and running the bench tests that prove a system meets its requirements * Demonstrated ownership of a complex subsystem from algorithm through released product, including evolving a substantial codebase you did not write * Systems-level thinking, with the ability to set technical direction independently and work across engineering, clinical, quality, and regulatory teams, * Software development in a regulated medical device environment: IEC 62304, ISO 14971, design controls, and documentation that has supported an FDA submission or audit * Familiarity, knowledge and/or experience with 21 CFR 820, Quality Management System Regulation, preferred/desired * Surgical navigation, robotic surgery, or another image-guided or safety-critical autonomous system, including patient-to-image registration * Robot integration: hand-eye calibration, forward and inverse kinematics, impedance or force control, vendor real-time control APIs or ROS, and functional safety * Cameras and depth sensors (stereo, structured light, time-of-flight), including vendor SDKs and driver-level integration * Closed-loop physiological control or biosignal processing, such as EMG, evoked potentials, or adaptive stimulation * GPU-accelerated computing such as CUDA, and ML inference frameworks such as TensorRT or ONNX Runtime * Familiarity with MRI-based anatomical models and coordinate spaces, and with TMS or neuromodulation ## Description The product is composed of Magnus Cloud and Magnus Neuromodulation software. The cloud analyzes fMRI to identify a personalized target within the brain linked to patients' emotional state. This target is then passed on to the neuromodulation software which guides clinicians through treatment using SAINT. We are extending the neuromodulation software with real-time neuronavigation and autonomous treatment delivery: a system that tracks the patient's head live, localizes the treatment coil in the patient's own imaging space, and drives a robotic arm to place and hold the coil on target through a session. The concept is proven and the production implementation is well under way. As a Staff Software Engineer you will be its technical owner, working at the intersection of computer vision, real-time systems, and robotics, and accountable for it through to release. The scope runs from live tracking and coil localization through robotic integration, automated coil placement, automated motor threshold determination, and the safety architecture that makes autonomous operation acceptable. You will also own the bench testing that proves clinical accuracy under FDA design controls. We expect real depth in one of these areas, systems-level thinking across the rest, and the judgment to know when a result is solid enough to guide the treatment of a patient. This is a senior individual contributor role, based onsite in Burlingame, CA. Essential Responsibilities * Own the real-time neuro-navigation and autonomy stack end to end: architecture, algorithms, performance, and correctness * Develop, optimize, and validate live head tracking and coil localization, treating accuracy, robustness, and latency as one problem: computer vision, registration, sensor fusion, and filtering against system-level accuracy requirements * Design and optimize the C++ runtime connecting live sensor and robot data to CV and ML inference: low-latency data flow, concurrency, and GPU acceleration within a fixed latency budget * Own integration with the robotic arm: real-time control interfaces, hand-eye and tool-tip calibration, coordinate frames across imaging, tracking, robot, and patient, and motion constraints for work near a patient * Build the closed-loop autonomy that delivers treatment: automated coil placement on target and continuous motion compensation and drift correction through a session * Develop automated motor threshold determination: closed-loop control of stimulation intensity, detection of the motor response, and adaptive threshold estimation * Design the safety architecture for a system operating autonomously on a patient: interlocks, workspace and force limits, clinician override, and defined behavior on tracking loss or other degraded conditions * Apply machine learning to perception where it outperforms classical methods: semantic segmentation, landmark detection, and instrument detection, validated to clinical standards * Own sensor integration across cameras, depth sensors, and optical trackers: calibration, time synchronization, and error characterization * Own the accuracy and robustness test strategy and prove the system meets it: bench, phantom, and system-level accuracy testing with the fixtures and ground truth it depends on; identify failure modes; author the protocols and reports supporting design verification and FDA submissions * Own the medical device software lifecycle for these components: requirement definition, risk assessment, design documentation, and release * Own the interfaces to the upstream imaging pipeline that supplies anatomical models and targets, and to the clinical software that presents guidance * Set technical direction, mentor engineers, and work across clinical, scientific, quality, and regulatory teams * Work onsite in Burlingame, CA, with regular hands-on time on sensors, robot, and bench test fixtures ## Related Videos - [Robots are coming into the wild! Full-Stack Robotics Engineers, be ready!](https://www.wearedevelopers.com/videos/479-robots-are-coming-into-the-wild-full-stack-robotics-engineers-be-ready) - [Tour de Force: Open-Source LLM Inference Optimization from Simple to Sophisticated](https://www.wearedevelopers.com/videos/100099-tour-de-force-open-source-llm-inference-optimization-from-simple-to-sophisticated) - [Docker network without Docker](https://www.wearedevelopers.com/videos/1418-docker-network-without-docker) - [Developing an AI.SDK](https://www.wearedevelopers.com/videos/198-developing-an-ai-sdk) - [Efficient deployment and inference of GPU-accelerated LLMs​](https://www.wearedevelopers.com/videos/929-efficient-deployment-and-inference-of-gpu-accelerated-llms) - [Docker exec without Docker](https://www.wearedevelopers.com/videos/1094-docker-exec-without-docker) ## Related Articles - [How We Built a Worry-Free System That Runs for 10+ Years – And What We’d Do Again](https://www.wearedevelopers.com/magazine/751-how-we-built-a-worry-free-system-that-runs-for-10-years-and-what-we-d-do-again) - [Everything a Developer Needs to Know About MCP with Neo4j](https://www.wearedevelopers.com/magazine/604-everything-a-developer-needs-to-know-about-mcp-with-neo4j) - [How software is steering vehicle technology](https://www.wearedevelopers.com/magazine/515-how-software-is-steering-vehicle-technology) - [Highest Paying Tech Companies for Developers](https://www.wearedevelopers.com/magazine/220-highest-paying-tech-companies-for-developers) - [Graph and AI Trends 2026: Why Is AI Running but Not Yet Delivering?](https://www.wearedevelopers.com/magazine/680-graph-and-ai-trends-2026-why-is-ai-running-but-not-yet-delivering) - [Best US AI Conferences for CTOs in 2026: Build vs. Buy, Vendor Evaluation, and Peer Intelligence](https://www.wearedevelopers.com/magazine/736-best-us-ai-conferences-for-ctos-in-2026-build-vs-buy-vendor-evaluation-and-peer-intelligence)