AI Systems

<h3><strong>About Unconventional</strong></h3> <p>Since 2022, AI has entered the mainstream, reshaping entire industries from education and software development to fundamental consumer behaviors. This revolution has created an unprecedented demand for computation - a demand that is now fundamentally limited by energy, not just in the datacenter, but at a global scale.</p> <p>At Unconventional, our mission is to solve this. We are rethinking computing from the ground up to build a new foundation for AI that is 1000x more efficient. We're doing this by exploiting the rich physics of semiconductors, mapping neural networks directly to the device physics rather than relying on layers of inefficient abstraction.</p> <h3><strong>The Role</strong></h3> <p>As a Member of Technical Staff, AI Systems, you will develop state-of-the-art architectural components, write their bespoke implementations for our unconventional software framework, and map them efficiently down to the physical silicon. You are critical to preparing our software stack for upcoming tapeouts by acting as the bridge between model architecture and physical compute.</p> <h3><strong>What You'll Do</strong></h3> <ul> <li><strong>AI Architectural Modeling:</strong> Co-design and evaluate next-generation AI models (e.g, transformers, diffusion, flow, and energy-based models). You will collaborate closely across the team to combine, modify, and implement core modeling components, including both conventional (e.g., attention, normalization, Mixture-of-Experts, FFNs) and unconventional components. You will ensure that they function optimally across our novel compute substrates.</li> <li><strong>Performance Modeling &amp; Scaling:</strong> Establish and test scaling laws specific to our novel&nbsp; hardware. Develop rigorous performance models to evaluate compute vs. memory trade-offs</li> <li><strong>Advanced Mapping &amp; Partitioning:</strong> Drive the partitioning and mapping of complex AI models down to hardware. &nbsp;Apply and invent advanced optimization strategies from first principles, including custom quantization schemes, sparsity/pruning, and distillation to fit the physical constraints of our substrates.</li> <li><strong>GPU Optimization &amp; Kernel Development:</strong> Develop and optimize GPU kernels using low-level programming models like <strong>CUDA</strong>, <strong>Triton</strong>, or <strong>CUTLASS</strong>. Profile and debug complex ML codebases to resolve performance bottlenecks (training and inference).</li> <li><strong>Cross-Functional Collaboration:</strong> Act as a translator, discussing algorithmic trade-offs with theorists and converting model requirements into concrete specifications for infrastructure and hardware engineering teams.</li> </ul> <p>&nbsp;</p> <h3><strong>Minimum Qualifications</strong></h3> <ul> <li><strong>Education:</strong> An <strong>MS/PhD or equivalent research/project experience</strong> in a quantitative field such as AI/Machine Learning, Computer Science, Physics, Electrical Engineering, or Applied Math.</li> <li><strong>Experience: </strong>Deep, practical understanding of the <strong>modern AI/ML stack</strong> and optimized compilation and execution of algorithms on modern GPU systems. Proven experience in profiling, identifying, and resolving performance bottlenecks in complex ML codebases.</li> <li><strong>Systems Fluency: </strong>Demonstrated ability to map state-of-the-art AI model architectures (e.g., Transformers, Mixture of Experts, diffusion models) to system performance implications and apply advanced efficiency techniques such as sparsity, quantization, and distillation.</li> <li><strong>Software Development: </strong>Deep experience with <strong>PyTorch</strong>, including its internals, torch.compile, and distributed data parallel (DDP) / fully sharded data parallel (FSDP) libraries.</li> </ul> <h3><strong>Preferred Qualifications (Nice to Have)</strong></h3> <ul> <li><strong>Unconventional Co-Design:</strong> A forward-looking perspective on co-designing algorithms for <strong>unconventional computing paradigms</strong> that map closely to the physics of underlying systems.</li> <li><strong>Next-Gen Efficiency:</strong> Theoretical or research experience in advanced approximation/compression techniques beyond standard quantization.</li> </ul> <h3><strong>Why Join Us?</strong></h3> <ul> <li><strong>The Mission:</strong> Redefine computing for the next 50 years by solving the fundamental energy limitation of AI at a global scale.</li> <li><strong>The Impact:</strong> Shape the company's future as a foundational team member. Enjoy massive ownership and an outsized opportunity to drive change.</li> <li><strong>The Perks:</strong> A comprehensive package including best-in-class health benefits, 401k matching, truly unlimited PTO, and complimentary meals in our Palo Alto office.</li> </ul>