Edge & Physical AI SW Product Manager/Director

NXP Semiconductors

11d

About The Position

This role is responsible for defining and managing the software product roadmap for NPU (Neural Processing Unit) software, covering the full stack from AI/ML compiler pipelines and optimization tools to runtime, SDKs, and developer APIs. The position involves creating a scalable platform to support various model architectures, engaging with Go-To-Market teams, enabling system-level solutions, and defining an integrated compiler strategy. The role requires continuous monitoring of AI market trends, driving integration strategies with existing developer environments, and owning the end-to-end developer experience. It also includes managing framework and ISV partner relationships, running beta programs, defining software policies, and performing competitive benchmarking. A working knowledge of firmware and driver interfaces is also necessary, along with developing system software solutions for industry events.

Requirements

8+ years of AI product management experience in software, developer tools, or AI/ML infrastructure, with direct ownership of a software product roadmap and developer-facing API surface.
Strong understanding of the inference model deployment pipeline—training frameworks (PyTorch, TensorFlow), model exchange formats (ONNX, TFLite), and inference runtime environments.
Familiarity with compiler toolchain concepts: graph optimization, operator fusion, quantization, and code generation—sufficient to write actionable requirements and evaluate engineering tradeoffs.
Experience defining and tracking developer adoption metrics—SDK downloads, model compilations, active deployments, and support ticket analysis.
Track record of driving open-source or partner ecosystem strategy—upstream contributions, ISV enablement, or developer community programs.
Strong prioritization discipline—experience managing a large backlog across multiple engineering teams with competing priorities.
Excellent technical communication skills; able to write a developer-facing API specification and an executive strategy memo in the same week.

Nice To Haves

Direct experience with MLIR, TVM, ONNX Runtime, or other open-source ML compiler frameworks.
Hands-on experience with model quantization—INT8/INT4/FP16, calibration workflows, and accuracy validation.
Background in NPU, GPU, or inference accelerator software stacks—Qualcomm SNPE/QNN, Hailo Dataflow Compiler, TensorRT, Core ML, or equivalent.
Experience with heterogeneous execution environments—managing workload dispatch across CPU, GPU, and dedicated accelerators.
Familiarity with embedded Linux, Android, or RTOS software environments for edge deployment.
Understanding of automotive software standards, including AUTOSAR and ASPICE, as well as functional safety and security, is a plus for MCU and MPU alignment.
MBA or advanced degree in computer science, electrical engineering, or a related discipline.

Responsibilities

Define and maintain the NPU software product roadmap spanning the full stack: AI/ML compiler pipeline, AI/ML tools for optimization (conversion, quantization, pruning), runtime and execution engine, SDK and developer APIs, model optimization, workflow and repository, and existing eIQ platform integration.
Own and create a scalable platform/toolkit to support newer model architectures (CNNs, LLMs, ViTs, VLMs) and enable a path to Day 0 model support in the long run.
Maintain familiarity with supported architectures and operators, align customer expectations with available model support, and work with Engineering to close identified gaps.
Engage closely with GTM to help prioritize industry verticals, segments, and use cases to create an enablement strategy for all business units.
Enable system-level solutions through pre-sales and sales strategy and support.
Define an integrated compiler strategy and transition all NXP AI products to a unified AI workflow across business units.
Continuously monitor AI market developments and trends (compilers, operators, frameworks, models, microservices, skills, and agents) and build or refine NXP’s AI strategy.
Drive the eIQ integration strategy—integrate the current- and next-generation Neutron compiler and SDK within NXP's eIQ developer environment and define NPU support for the existing MPU/MCU developer base. Create a cradle-to-grave plan for features and components during eIQ integration.
Develop use-case-driven software enablement and help prioritize, reset, and reconcile existing features to deliver a holistic user experience.
Own the upstream open-source compiler integration strategy—define NXP's participation and contribution roadmap for Apache TVM, MLIR/LLVM, ONNX Runtime, and PyTorch/TensorFlow export paths.
Define and maintain the model zoo or an equivalent—curate, compile, quantize, and benchmark frontier models (LLaMA variants, YOLO generations, Whisper, Phi, Gemma, custom vision transformers) on Ara NPU and publish validated results.
Own the developer experience end-to-end: SDK installation, model import workflow, compilation, profiling, debugging, and deployment—define and track developer adoption funnel metrics.
Manage framework and ISV partner relationships—PyTorch, TensorFlow, Hugging Face, ONNX Runtime, and key vertical AI software partners; drive NPU support contributions and co-marketing. Support external software integrations and licensing partnerships.
Run beta and early-access programs with strategic developer customers; translate feedback into prioritized engineering requirements.
Define software versioning policy, backward compatibility commitments, and long-term support (LTS) release strategy.
Own competitive benchmarking of model coverage against Qualcomm AI Hub, DeepX, Axelera Metis SDK, and Arm NN/ExecuTorch while preserving software ease of use.
Maintain a working knowledge of firmware and PCIe and USB drivers.
Develop end-to-end system software solutions and demonstration platforms for hardware and software developers at industry events and conferences.