Computer Vision Engineer III, Senior (Rendering and Simulation)

Lockheed Martin•Littleton, CO

3d•Onsite

About The Position

Lockheed Martin Space is seeking a motivated engineer to join their team and take ownership of their in-house rendering engine, which is the core tool that generates synthetic, radiometrically accurate scenes for testing and validating computer vision solutions. In this role, the engineer will extend and enhance the engine to meet the demanding performance, fidelity, and hardware-constraint requirements of space-borne platforms. The team specializes in classical computer-vision algorithms like feature detection, stereo matching, and optical flow, while also integrating modern AI techniques such as semantic segmentation and learned feature representations. Every solution is engineered with the strict hardware constraints of resource-limited processors in mind, including the ultra-low-power, radiation-hardened CPUs typical of space-borne systems.

Requirements

Proficiency in modern C++ (C++14/17/20)
Demonstrated ability to develop CUDA kernels. Hands-on experience writing and debugging CUDA (or equivalent) kernels
Experience with implementing computer graphics algorithms
Familiarity with GPU architectures, GPU memory hierarchies, and trade-offs between compute and memory bandwidth
Experience with GPU performance-profiling tools like NVIDIA Nsight, visual profiler, or something similar
Experience with github or gitlab
U.S. Citizen
Eligible to obtain, and maintain, Top Secret with adjudicated SCI Eligibility (i.e. TS/SCI) for long term employment on this program
Bachelor’s degree in Computer Science, Electrical Engineering, Mathematics, or a related technical field (or equivalent professional experience) with 3+ years of software development experience

Nice To Haves

Knowledge of physically-based rendering techniques, material models, and radiometric accuracy
Experience with ray-tracing engines
Experience with rendering pipelines and generating synthetic imagery
Familiarity with camera optics and sensor modeling
Experience with containerization (Docker)

Responsibilities

Quickly get up to speed with the current codebase, architecture, and data pipelines, identifying improvement opportunities and documenting key components.
Accelerate rendering with GPU compute. Design and implement CUDA-based kernels to achieve real-time or faster than real time performance for ray-traced scene generation.
Implement high fidelity camera models. Collaborate with optical engineers to translate sensor characteristics into accurate simulation environments.
Work with and update the C++ API. The API exposes the engine's core functionality. This needs to be maintained, updated, and documented for the use of downstream simulation pipelines.
Integrate and validate. Work hand-in-hand with algorithm developers to embed the rendering engine into their testing workflows, ensuring generated data meets the strict accuracy and performance requirements of space-borne platforms.
Define test cases, compare simulation results against analytical/ground-truth data, document V&V evidence, and support formal review packages.
Debug computer vision algorithm performance in the simulation by debugging the underlying code and using visualizations to diagnose issues.
Collaborate with embedded-hardware, optics, and systems teams to ensure the simulation reflects actual hardware behavior and integrates smoothly with downstream toolchains.