AI Systems, Training

Unconventional AI; Palo Alto (CA); On-site; USD 210,000 - 320,000; Full time

Benefits offered by this job

Health benefits

401k matching

Unlimited PTO

Complimentary meals in Palo Alto

Job summary

Unconventional AI in Palo Alto seeks a senior ML systems engineer to design and build the next-generation training platform. You will co-design and implement training ecosystems alongside novel AI models and hardware platforms that push the boundaries of physics-based compute.

You will optimize multi-node distributed training, implement elastic sharding, and develop robust data streaming and checkpointing, with a focus on CUDA/Triton kernels and MFU benchmarking.

Qualifications

• MS/PhD or equivalent in AI/ML, CS, Physics, Electrical Engineering, or Applied Math.
• Experience mapping state-of-the-art AI model architectures to system performance with distributed training and parallelism.
• Proven track record implementing production-grade training frameworks (Megatron-LM, DeepSpeed, Ray, PyTorch Lightning).

Responsibilities

• Build and maintain highly optimized, model-specific training stacks for Gen Vision, language, and world models.
• Design and scale multi-node distributed training with elastic sharding and robust data streaming pipelines.
• Develop and optimize kernels using CUDA and Triton; design benchmarking for MFU, memory bandwidth, and convergence.
• Translate model requirements into concrete infrastructure and hardware specs for cross-functional teams.

Job description

About Unconventional

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.

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.

The Role

You will be a key contributor to our training ecosystem. Your goal is to build the next-generation ML model training platform tailored for a world where compute is no longer constrained by the digital abstraction.

You will co-design and implement training systems alongside novel AI models and hardware platforms that push the boundaries of physics-based compute.

What You’ll Do

• The Model Architectures: Build and maintain highly optimized, model-specific training stacks specifically tuned for state-of-the-art generative vision, language, and world models.
• The Training Infrastructure: Design and scale multi-node distributed training systems, implementing elastic sharding and robust data streaming pipelines for fast, large-scale iteration. Implement and robust model checkpointing and recovery mechanisms.
• Optimization & Benchmarking: Develop and optimize kernels using low-level programming models like CUDA and Triton. Design rigorous benchmarking suites to track Model Flops Utilization (MFU), memory bandwidth, and convergence stability.
• Cross-Functional Collaboration: Act as a translator, discussing algorithmic trade-offs with theorists and converting model requirements into concrete specifications for infrastructure and hardware engineering teams.

Minimum Qualifications

• Education: An MS/PhD or equivalent research/project experience in a quantitative field such as AI/Machine Learning, Computer Science, Physics, Electrical Engineering, or Applied Math.
• Experience: Veteran of the modern ML software stack. Demonstrated ability to map state-of-the-art AI model architectures (e.g., transformers, Mixture of Experts, diffusion models) to system performance implication. Deep expertise in how models are partitioned across a cluster, with a mastery of communication primitives, and parallelism strategies.
• Software Development: Proven track record of implementing, debugging, and maintaining production-grade training frameworks—such as Megatron-LM, DeepSpeed, Ray, PyTorch Lightning—turning raw compute into a reliable model-building factory.

Preferred Qualifications (Nice To Have)

• Unconventional Co-Design: A forward-looking perspective on co-designing algorithms for unconventional computing paradigms that map closely to the physics of underlying systems.

Why Join Us?

• The Mission: Redefine computing for the next 50 years by solving the fundamental energy limitation of AI at a global scale.
• The Impact: Shape the company's future as a foundational team member. Enjoy massive ownership and an outsized opportunity to drive change.
• The Perks: A comprehensive package including best-in-class health benefits, 401k matching, truly unlimited PTO, and complimentary meals in our Palo Alto office.