Senior Staff Software Engineer, AI Model Enablement & Software Tools

Qualcomm • San Diego, CA

1d • $194,400 - $291,600

About The Position

As a Senior Staff Software Engineer in Qualcomm’s AI Stack SDK Software team, you will provide technical leadership in bringing state-of-the-art AI models to Snapdragon platforms using Qualcomm’s AI software stack and model-enablement frameworks. You will lead the analysis, adaptation, optimization, quantization, compilation, and execution of LLM, LVM, VLM, and multimodal architectures, working with AI Research to understand model internals and address operator- and graph-level gaps with Qualcomm AI accelerator backends.

The role spans model architecture, machine learning algorithms, graph transformations, quantization, compiler and runtime technologies, and hardware-aware optimization. You will set technical direction and improve the AI software stack through reusable model adaptations, quantization recipes, graph optimizations, compilation flows, and execution interfaces. You will collaborate with research, hardware and software engineering, product, program management, QA, and customer engineering teams to take model architectures from initial analysis through production-quality execution on Qualcomm devices.

This is a hands-on leadership role. You will lead and manage a small team while contributing as an individual engineer, driving architecture and implementation, setting engineering priorities, mentoring team members, and owning delivery across multiple model-enablement initiatives.

Requirements

• Bachelor’s degree in Computer Science, Computer Engineering, Electrical Engineering, or a related field and 8+ years of experience in software design, development, and delivery; or a master’s degree or PhD in a related field and 6+ years of such experience.
• At least 5 years of hands-on experience in AI/ML software development, model inference, model optimization, compiler technologies, or related systems software.
• Strong knowledge of modern AI model architectures, including transformer-based LLMs, vision models, and multimodal or vision-language models.
• Deep understanding of transformer internals, including attention, positional encoding, KV cache, prefill and decode execution, normalization, feed-forward networks, and emerging architecture patterns.
• Strong understanding of quantization and calibration, including PTQ, QAT, mixed precision, weight and activation quantization, encoding generation, and accuracy-performance tradeoffs.
• Strong knowledge of computational graphs, graph transformations, operator decomposition, graph lowering, layout and shape transformations, and backend-aware optimization.
• Excellent software architecture and system-design skills, with experience designing scalable, modular, extensible, maintainable frameworks, APIs, and production software systems.
• Proficiency in Python and C/C++ for production-quality software development.
• Strong experience with PyTorch and ONNX model representations and tooling.
• Proficiency with agentic coding tools such as Claude Code, OpenAI Codex, or equivalent for codebase exploration, design, implementation, testing, debugging, and code review, with engineering judgment to validate AI-generated solutions.
• Ability to investigate unfamiliar model architectures, identify software and backend gaps, and develop scalable technical solutions.
• Strong debugging and root-cause-analysis skills across model, framework, compiler, runtime, and hardware boundaries.
• Technical leadership experience, including leading a small team, mentoring engineers, prioritizing work, and driving complex projects end-to-end.
• Ability to work independently and collaborate with research, hardware, software, product, QA, and customer-facing organizations.

Nice To Haves

• Experience bringing up LLM, LVM, VLM, or multimodal models on embedded, edge, mobile, or accelerator-based platforms.
• Experience designing model-enablement pipelines or frameworks for model loading, re-authoring, quantization, export, compilation, deployment, and execution.
• Experience with AIMET or equivalent model quantization and optimization frameworks.
• Experience with Hugging Face Transformers, model repositories, tokenizers, processors, PEFT, and related ecosystem components.
• Knowledge of advanced quantization techniques such as blockwise or groupwise quantization, weight rotation, adaptive scaling, sequential MSE optimization, and quantizer-placement optimization.
• Experience with graph optimization and compiler technologies such as ONNX Runtime, ExecuTorch, TFLite/LiteRT, MLIR, or equivalent frameworks.
• Experience with LoRA, MoE, speculative decoding, hybrid attention architectures, state-space or recurrent model components, and vision encoder/decoder models.
• Understanding of accelerator architecture, operator support, memory hierarchy, data movement, precision support, and hardware/software performance tradeoffs.
• Experience with Android and/or RTOS environments such as QNX.
• Experience with Linux development, CMake-based build environments, agile software development, and Git-based source control.
• Experience interacting with senior leadership and communicating technical strategy, execution status, and project risks.
• Previous experience leading and mentoring a globally distributed engineering team.
• User-level or development experience with commercial AI software stacks and model-enablement frameworks, including model conversion, optimization, compilation, deployment, and runtime tools.
• Exposure to Snapdragon SoCs and AI accelerators such as NPUs or HTPs.

Responsibilities

• Lead bring-up and enablement of new LLM, LVM, VLM, and multimodal model architectures on Snapdragon platforms.
• Partner with AI Research teams to understand model architecture, tensor flows, attention and cache behavior, model inputs and outputs, and performance-sensitive computation patterns.
• Analyze model graphs and identify operator-level, graph-level, precision, memory, and runtime gaps against Qualcomm AI accelerator backends.
• Own architecture and design of scalable model-enablement solutions, including component boundaries, APIs, stage contracts, configuration schemas, artifact interfaces, extensibility mechanisms, and integrations across model, quantization, compiler, and runtime systems.
• Define scalable approaches for model re-authoring, graph adaptation and lowering, decomposition, fusion, layout transformation, and backend-specific optimization.
• Architect reusable model-enablement solutions and integrate them into Qualcomm’s AI software stack and model-enablement frameworks.
• Lead design of declarative, extensible pipeline capabilities, including reusable stages, model mappings, quantization recipes, configuration interfaces, artifact handoffs, caching, validation, and execution workflows.
• Develop and optimize model preparation, quantization, calibration, export, compilation, and on-device execution flows.
• Apply quantization techniques including PTQ, QAT, mixed precision, weight and activation quantization, blockwise or groupwise quantization, calibration, and accuracy-recovery optimization.
• Evaluate accuracy, latency, throughput, memory usage, model size, and power efficiency across floating-point, prepared, quantized, compiled, and on-device execution checkpoints.
• Guide enablement of advanced Generative AI features such as KV-cache optimization, prefill and decode graphs, LoRA, Mixture-of-Experts, speculative decoding, multimodal processing, and vision encoders.
• Debug issues across model architecture, PyTorch and ONNX graphs, quantization, compiler, runtime, operating system, and hardware layers.
• Establish technical plans, divide work across the team, manage execution risks, and drive model-enablement initiatives from investigation through production delivery.
• Participate in and lead architecture, design, code, performance, and accuracy reviews.
• Mentor engineers, develop technical expertise within the team, and establish practices for maintainable, production-quality AI software.
• Communicate technical strategy, project status, dependencies, and risks to engineering leadership and cross-functional stakeholders.

Benefits

• Competitive annual discretionary bonus program.
• Opportunity for annual RSU grants.
• Competitive benefits package designed to support success at work, at home, and at play.

What This Job Offers

• Job Type: Full-time
• Career Level: Senior