Toronto, Ontario, Canada
I am a multi-disciplinary engineering leader specializing in managing the design and deployment of high-performance computing (HPC) and artificial intelligence (AI) applications. My technical approach to design in the age of AI is to double-down on engineering fundamentals: Problem solving and experimentation using effective workflows executed by AI agents with minimal human interventions required. My business leadership style focuses on maintaining alignment between the senior leadership and team members, and fostering trust to meet objectives. I advocate for practical software design processes, continuous improvement initiatives and engineering best practices to outperform competitors.
- Software stack Performance for Meta's Training and Inference Accelerators (MTIA). - Designing agentic engineering workflows for teams to deliver domain expertise in an AI-native environment at scale. https://ai.meta.com/blog/meta-mtia-scale-ai-chips-for-billions/
Untether AI designs and manufacturers neural network accelerators using at-memory computing for AI workloads ideal for Autonomous Driving and vision workloads. Reporting to the Vice-President of Software Engineering, I execute in a multi-disciplinary role providing planning, review and technical assistance to software, firmware and hardware teams. - responsible for the overall system performance, helping teams optimize for compute and memory efficiency, both on the host and on the accelerator. - Work directly with the Product Team to assist in customer SOW review and customer scaleout planning. - develop requirements and architectures between chip, board, firmware and software teams - directly support the runtime team in optimizing the kernel and userspace drivers - collaborate with the Dev Infrastructure Lead in continuous improvement initiatives for DevOps
Technical Advisor to the CEO/CTO. - advise on investor pitches and industry partnerships, market strategy, leadership mentoring and coaching. Co-founder for Multiverse product - licensed mixed-signal PCB designs to Aviate Audio for product inception. - licensed various software IP to Aviate Audio for commercial use. - designed and implemented cryptographic security features for hardware/software authentication. - wrote UX applications and device firmware for product MVP.
Algolux was acquired by Torc Robotics (Daimler). The Automotive Perception and Deep Learning team is responsible for porting and implementing Algolux's award winning CNN-based vision stack on real-time hardware devices for automotive OEM and Tier 1 customers in autonomous driving and ADAS applications. As R&D Manager, reporting to the CTO, I was responsible for - 7 direct reports: hiring within the team, performance and salary reviews, career development planning and training - work with Project Engineering on estimates, customer SOWs, bids and proposals. - work with CTO on supporting technology roadmap - Software Architecture for Eos product, ownership of software design life-cycle processes and department design processes (Internal and Aspice) The Embedded Deep Learning Software Team is responsible for: - converting CNN models from our Research team (TensorFlow, PyTorch, Caffe) to ONNX and TensorRT - converting computer vision and image processing algorithms from Python to C++ and CUDA - optimizing real-time performance across multiple hardware platforms including NVIDIA GPUs, NVIDIA Drive/Jetson, Automotive vision accelerators (Renesas, FPGAs) - neural networks for detection, segmentation and other autonomous driving features
Reporting to the Director of SW Engineering, I was responsible for the Architecture of our embedded high-performance computing systems, driving our Software Concept Study and Prototyping pipelines, and providing coaching and mentoring to a department of over 60 software engineers.
The Advanced Processing group is a cross-functional team of algorithm developers, data scientists and implementation specialists focusing on machine vision, AI and image processing. These specialists transform academic models into practical solutions by partitioning and deploying on to various hardware accelerators including SIMD processors, FPGAs and GPUs.
A member of the Software Management Team (3 people) providing oversight to a department of over 50 engineers with regards to: - staff resource assignment - project scheduling and estimates - continuous improvement on departmental practices - formal performance reviews - interviewing/recruiting candidates to meet business growth Technical Leader for High Performance Computing (HPC) projects. This involves deploying digital signal and data processing across a heterogeneous collection of acceleration platforms including GPUs, VLIW / SIMD accelerators and FPGAs. I rely on my broad experience (summarized below) in hardware and software to carry out the following tasks: - directly lead a cross functional team of software and FPGA engineers working in the areas of accelerated computing, machine vision and neural networks - work with Senior Management to develop and maintain a technology roadmap - carry out research and concept work for upcoming technology platforms and applications - participate in design reviews and NPI stage gates - gather and supply key performance indicator metrics in reports to Sr. management.