Candidate Availability
Required and preferred rules are kept separate and reflect the wording in the original posting.
About the Role
You will own technical strategy for a major area of the application layer, turning research, open-source projects, and product requirements into roadmaps for model deployment, pipeline development, and application integration. You will design reusable architectures, optimize pipeline performance across hardware platforms, develop profiling tools, and resolve complex compilation and deployment challenges. You will also guide technical direction, mentor engineers, document best practices, and represent technical work to stakeholders and customers.
Requirements
- BS/MS in Computer Science, Electrical Engineering, or equivalent experience.
- Typically 8+ years of hands-on experience in the semiconductor and/or AI industry.
- Expertise with edge deployment frameworks such as OpenVINO or TensorRT.
- Experience with GPU computing APIs such as OpenCL and Vulkan.
- Experience applying model optimization techniques including quantization, compression, and pruning.
- Expert AI application development skills using Python and ML libraries such as PyTorch and TensorFlow.
- Expert proficiency in Python and C++.
- Experience architecting, building, and deploying end-to-end pipelines with quantized models at scale.
- Experience creating reusable frameworks.
- Experience training models using transfer learning.
- Strong technical judgment and ability to navigate ambiguity.
- Strong communication and influencing skills with senior technical and business stakeholders.
- Experience mentoring and developing engineers.
- Proficient Linux skills.
- Practical agile development knowledge using Jira, Git, and GitHub.
Responsibilities
- Own the end-to-end technical strategy for a major area of the Voyager SDK application layer.
- Translate research papers, open-source repositories, and product requirements into a roadmap for model deployment, pipeline development, and application integration.
- Set technical direction for image pre-processing and post-processing operators and decoders.
- Create reusable architectures and partner with the compiler team to resolve complex compilation issues.
- Integrate industry-standard model frameworks into the SDK.
- Define reusable libraries and architecture for low-code and no-code deployment of models and datasets.
- Define metadata representations for common model types, including bounding boxes and keypoints.
- Own libraries for evaluating deployed-model accuracy and rendering inference results on hardware.
- Take accountability for end-to-end pipeline latency and throughput across supported hardware platforms.
- Diagnose bottlenecks and build profiling tools and frameworks for performance analysis.
- Represent technical capabilities to business stakeholders, customers, and the industry.
- Mentor engineers and drive documentation standards and best practices.
Benefits
- Pension plan.
- Extensive employee insurances.
- Option to receive company shares.