Software Architect

Job Title: Edge AI Software Architect

Job Description

This high-impact, highly visible role sits at the intersection of silicon, software, and AI. You will own the end-to-end software architecture of an Edge AI platform that enables customers to deploy AI models from prototype to production on heterogeneous SoCs with minimal friction. Working closely with silicon architects, runtime engineers, compiler teams, and customers, you will translate real-world deployment challenges into scalable platform solutions and drive system-level design across inference engines, runtime orchestration, and heterogeneous hardware targets including CPU, GPU, NPU, and DSP.

Responsibilities

• Define and own the Edge AI software architecture spanning multiple inference engines such as ONNX Runtime, LiteRT, and ExecuTorch, as well as the overall runtime orchestration layer.
• Design a unified inference runtime that supports current and future frameworks, ensuring hardware-accelerated execution paths for NPU, DSP, GPU, and multi-core CPU targets.
• Ensure architecture portability and consistency across MCU, MPU, and accelerator-based SoC families, enabling a scalable Edge AI platform.
• Drive integration of inference engines into system software, including OS and RTOS drivers, middleware, and hardware abstraction layers, to enable seamless model deployment from prototype to production.
• Shape AI toolchain requirements for model conversion, compilation, quantization, profiling, and optimization, and collaborate closely with compiler and SDK teams to address gaps and improve workflows.
• Architect and refine model optimization and compilation pipelines, including quantization, pruning, operator fusion, and graph optimization, to meet power, latency, and performance targets.
• Champion developer experience by reducing friction in the development process, improving debugging and observability, and ensuring a smooth customer bring-up journey for AI applications.
• Provide technical leadership as the bridge across silicon, runtime, compiler, and application teams, driving alignment, trade-off decisions, and coherent platform strategy.
• Evaluate emerging Edge AI technologies such as on-device large language models, agentic AI, and hardware–software co-design approaches, and guide their adoption into the platform roadmap.
• Represent the platform architecture in customer engagements, clearly explaining design decisions and constraints, and influence outcomes without direct authority across distributed engineering teams.
• Review and understand software implementations and code architecture, providing guidance to engineering teams to ensure robust, maintainable, and production-ready embedded software.
• Collaborate with internal and external stakeholders to influence open-source inference frameworks, ML compilers, and industry standards from a position of silicon and platform expertise.

Essential Skills

• Minimum 10 years of experience in embedded systems, system software, or AI/ML platform engineering.
• Minimum 2–3 years of hands-on software architecture experience, preferably in complex embedded software environments.
• Proven experience architecting software solutions across heterogeneous compute platforms, with demonstrated work on at least two architectures such as CPU plus NPU, DSP, or GPU.
• Expert-level C++ programming skills with deep knowledge of modern C++ standards (C++14, C++17, C++20); strong C knowledge combined with advanced C++ expertise is mandatory.
• Deep knowledge of AI inference frameworks used in embedded environments, including ONNX Runtime, TensorFlow Lite or LiteRT, PyTorch or ExecuTorch, or equivalent solutions.
• Strong understanding of SoC architecture, including memory hierarchy, DMA, power domains, and hardware accelerator integration for AI workloads.
• Ability to review, understand, and provide guidance on software implementations and code architecture in production embedded systems.
• Excellent communication skills, with the ability to present complex architecture decisions to both technical and executive audiences.
• Bachelor’s or Master’s degree in Computer Science, Electrical Engineering, or a related field, or equivalent practical experience.

Additional Skills & Qualifications

• Experience with model optimization techniques such as quantization (including INT8 and INT4), pruning, operator fusion, and graph optimization for embedded AI deployments.
• Experience working with major SoC ecosystems and their AI acceleration stacks, such as those from Qualcomm, Arm, or similar vendors.
• Familiarity with RTOS environments like FreeRTOS and Zephyr, as well as bare-metal approaches to embedded AI deployment.
• Knowledge of compiler backends and ML compiler technologies, including LLVM, MLIR dialects, TVM relay, or custom NPU compilers.
• Experience with on-device LLM inference solutions such as , MLC-LLM, or similar toolchains, and with multimodal AI pipelines.
• Experience in automotive, industrial, or other safety-critical AI deployment domains, with familiarity with standards such as ISO and IEC relevant to functional safety.
• Open-source contributions to inference frameworks, ML compilers, embedded AI toolchains, or related Edge AI software projects.
• PhD in Machine Learning, Computer Architecture, or a closely related field is a strong plus.
• Interest in and understanding of Edge AI, embedded AI, heterogeneous compute, real-time AI, computer vision, signal processing, power efficiency, and latency optimization.
• Motivation to work on AI platforms that ship in hundreds of millions of devices across automotive, industrial, IoT, and consumer markets, and to collaborate with world-class silicon architects, compiler engineers, and AI researchers.

Work Environment

This role is based in Ottawa in a hybrid work model, with three days per week onsite and the remaining time available for remote work. The position is initially structured on a yearly basis with potential for long-term engagement depending on performance and business needs. You will operate in a cutting-edge technical environment focused on Edge AI, working with technologies such as ONNX Runtime, TensorFlow Lite, LiteRT, ExecuTorch, PyTorch, MLIR, TVM, LLVM, and custom AI compilers, as well as heterogeneous hardware including NPU, DSP, GPU, MCU, MPU, and complex SoC architectures. The culture emphasizes a global engineering mindset, collaboration across silicon, software, and AI research teams, and flexible work arrangements. Competitive compensation and benefits, opportunities to influence open-source frameworks and industry standards, and the chance to shape production-ready Edge AI platforms used in automotive, industrial, IoT, and consumer devices make this environment both technically challenging and highly rewarding.

Job Type & Location

This is a Contract position based out of Ottawa, ON.

Pay and Benefits

The pay range for this position is $65.00 - $100.00/hr.

Workplace Type

This is a hybrid position in Ottawa,ON.

À propos d'Actalent

Actalent est un leader mondial dans les services d’ingénierie et de sciences ainsi que dans les solutions de talents. Nous aidons des entreprises visionnaires à faire progresser leurs initiatives d’ingénierie et de science grâce à l’accès à des experts spécialisés qui favorisent l’échelle, l’innovation et la rapidité de mise sur le marché. Avec un réseau de près de 20 000 consultants et 5 000 clients à travers les États-Unis, le Canada, l’Asie et l’Europe, Actalent dessert de nombreuses entreprises du Fortune 500. Nous sommes fiers d’être l’une des 500 meilleures firmes de conception de l’Engineering News-Record (ENR) pour nos services de conception en ingénierie et un lauréat du prix ClearlyRated Best of Staffing tant pour le service client que pour les talents.

Actalent est un employeur souscrivant au principe de l’égalité des chances et accepte toutes les candidatures sans tenir compte de la race, du sexe, de l’âge, de la couleur, de la religion, des origines nationales, du statut d’ancien combattant, d’un handicap, de l’orientation sexuelle, de l’identité sexuelle, des renseignements génétiques ou de toute autre caractéristique protégée par la loi.

Si vous souhaitez faire une demande d’accommodement raisonnable, tel que la modification ou l’ajustement du processus de demande d’emploi ou d’entrevue à cause d’un handicap, veuillez envoyer un courriel à