3 months ago
Base Salary
$158k - $238k/yr
Responsibilities
- Convert, optimize, and deploy AI models from PyTorch and ONNX for efficient Snapdragon inference.
- Design and implement graph transformations, graph lowering, and optimization techniques in ONNX Runtime, ExecuTorch, and the Qualcomm AI Stack SDK.
- Apply quantization and performance optimization to improve latency, throughput, memory usage, and power efficiency.
- Develop and productize generative AI inference features involving transformer architectures, attention mechanisms, MoE, LoRA, and speculative decoding.
- Debug complex issues across models, runtimes, operating systems, compilers, and hardware while performing root-cause analysis.
- Design, implement, and deliver Qualcomm AI Stack SDK features and enhancements.
- Participate in design and code reviews and mentor junior engineers while driving execution across initiatives.
- Collaborate with ML Research, AI accelerator hardware/software, product management, program management, QA, and customer teams.
Requirements
- Bachelor’s degree and 6+ years of software design, development, and delivery experience for Staff level, or 8+ years for Sr. Staff level; alternatively, a master’s degree or PhD and 5+ years for Staff level or 7+ years for Sr. Staff level.
- At least 3 years of hands-on AI/ML software development experience focused on inference or model optimization.
- Strong understanding of AI/ML fundamentals, deep learning, inference pipelines, transformer architectures, and attention mechanisms.
- Proficiency in Python and C/C++ for production-quality software development.
- Experience with PyTorch and ONNX models and tooling.
- Ability to debug complex issues, perform root-cause analysis, and support high system reliability.
- Preferred qualifications include graph theory and graph optimization, compiler-style transformations, LLM/LVM/LMM inference pipelines, Hugging Face and PEFT, LoRA, MoE, Android or RTOS environments such as QNX, CMake, Git, embedded or system-level software, Qualcomm AI Stack SDKs such as QAIRT, QNN, and Genie, Snapdragon SoCs, NPU accelerators, and GenAI features.
- Preferred qualifications also include experience interacting with director-level and above leadership and mentoring junior engineers.
Benefits
- Full-time onsite position requiring five days per week in Qualcomm’s San Diego office.
- Competitive annual discretionary bonus program and opportunity for annual RSU grants.
- Competitive benefits package supporting employees at work, at home, and at play.
About Qualcomm
Inspired by 250 years of American ingenuity, we build technologies that expand what’s possible. From wireless to AI at scale and 6G, our innovations power progress worldwide.