Hardware/Software Co-design, Performance Modeling

Data movement bottlenecks are the single greatest limiting factor in developing novel hardware for machine learning. We use silicon photonics to break through these barriers, and we're re-architecting the entire AI hardware stack in order to provide multiple-order-of-magnitude improvements in bandwidth and performance. We are creating a hardware/software solution that will outperform the world's largest AI supercomputers.Luminous is backed by Bill Gates, Luke Nosek, Ali Partovi and Dara Khosrowshahi, among others.

About the job:

We are looking for an experienced and talented Performance Modeling Software Engineer to work in a fast paced engineering organization designing cutting-edge technologies in the AI HW acceleration space. In this role you will drive the development, architecture, analysis, qualification, and quantification of our system-level performance models, simulation software, and supporting tooling and infrastructure. The tools and processes developed by the team will be used to validate, verify, and explore cross-cutting, system-level design decisions across the various engineering domains at Luminous. You will be a foundational member of a small team tasked with building a platform to model and explore the hardware/software design space of Luminous Computing's first series of products.

What you will be doing:

• Build analytical performance models for deep-learning workloads executing on novel hardware system designs and integrate them into a sophisticated system-level simulation platform
• Model the performance of Luminous hardware systems with sufficient precision to influence and/or validate critical system architectural design decisions
• Work closely with HW design teams to implement behavioral performance models of AI HW accelerator system components (compute, memory, & interconnects) within our internal system-level architecture modeling and simulation framework
• Participate in identifying system and component-level bottlenecks that exist within and/or across the network, memory, and compute subsystem boundaries
• Propose new architectural hardware and software features to improve runtime of ML models of interest on current and future architectures
• Inform automatic parallelization techniques and optimized kernel generation within the compiler through performance modeling and iterative parameter sweeping over the parallelization strategy and network topology domains

Requirements

• 3+ year of experience co-designing and implementing hardware and software systems to accelerate machine learning and deep learning workloads
• BS/MS/PhD in Computer Science, Computer Engineering, or a related technical field

• Proficiency in modern programming languages (e.g. C++, Python, Haskell, etc.)

• Strong independent problem solving skills coupled with the ability to work collaboratively with a small software-engineering team working on mission-critical projects

Ways to stand out:

• Background in Computer architecture and familiarity with processor and system performance modeling
• Experience in performance analysis and workload profiling
• Understanding of modern deep neural network architectures (e.g. CNNs, Transformers, GNNs, Mixture-of-Experts, DLRM, etc.) and the algorithms that drive them (e.g convolutions, attention, back-propagation, etc.)
• Experience developing and deploying machine learning models in the datacenter (TensorFlow, PyTorch, MeshTensorflow, DeepSpeed, etc.)
• Experience optimizing software routines (i.e. kernels) using knowledge of the underlying low-level details of the target hardware architecture

Benefits

Why Luminous Computing?

• Our benefits are here to support you!
  • Competitive salary and benefits
  • 401(k) program with company match of 100% up to 4% of your annual salary
  • Comprehensive medical / dental / vision, STD/LTD, AD+D and Life Insurance
  • Flexible vacation and company paid holidays
  • Daily lunch, dinner, snacks, and drinks for on-site employees
  • 12 weeks of paid Parental Leave
  • Flexible Spending Account (FSA), Health Savings Account (HSA)
  • Pre-Tax Transportation and Commuter Benefits
  • Mental Health Benefits

• Were remote friendly
  • We believe that your value to the team can be provided at a Luminous Computing office or from the comfort of your home. Depending on your function, enjoy the flexibility of remote work and join our growing community of remote employees in the continental United States.

• Were vested
  • You wont just own your work here; youll have the potential to own equity in Luminous Computing. We are competing in a market that is projected to grow exponentially, which gives our company valuation room to grow at tremendous rates.

Luminous Computing provides equal employment opportunities (EEO) to all employees and applicants for employment without regard to race, color, religion, sex, national origin, age, disability, or genetics. In addition to federal law requirements, Luminous abides by applicable state and local laws governing nondiscrimination in employment. This policy applies to all terms and conditions of employment, including recruiting, hiring, placement, promotion, termination, layoff, recall, transfer, leaves of absence, compensation, and training.

Note to Recruitment Agencies: Luminous Computing does not accept unsolicited agency resumes. Furthermore, Luminous Computing does not pay placement fees for candidates submitted by any agency other than its approved partners.

Recommended Skills

• Algorithms
• Architecture
• Artificial Intelligence
• Artificial Neural Networks
• C++ (Programming Language)
• Computer Architectures