AI Systems, Model Optimization at unconventional

Department: Engineering

Setup: Remote
Location: Mountain View, California
Level: staff+
Posted: Jul 21, 2026

Description

Unconventional is hiring a Member of Technical Staff in AI Systems and Model Optimization to develop training techniques, optimization strategies, and infrastructure for running AI models efficiently on novel analog compute substrates. The role covers energy benchmarking, hardware-aware model mapping and partitioning, quantization-aware and noise-aware training, GPU kernel development, and collaboration between model architects and hardware engineering teams. Candidates need an MS/PhD or equivalent research or project experience in a quantitative field, practical GPU optimization experience, and deep PyTorch expertise.

Required Skills

PyTorch
CUDA
Triton
CUTLASS
Quantization-Aware Training
Sparsity
Distillation
Distributed Data Parallel

Additional Information

Minimum Education: MS/PhD or equivalent research or project experience in a quantitative field such as AI/Machine Learning, Computer Science, Physics, Electrical Engineering, or Applied Math.
Required Previous Experiences: Practical understanding of the modern AI/ML stack and optimized compilation and execution of algorithms on modern GPU systems.
Profiling, identifying, and resolving performance bottlenecks in complex machine learning codebases.
Deep experience with PyTorch, including its internals, torch.compile, and distributed data parallel or fully sharded data parallel libraries.
Preferred Qualifications: Experience with production-grade training frameworks such as Megatron-LM and DeepSpeed, and distributed training at scale.
Perspective on co-designing training systems for unconventional computing paradigms aligned with underlying system physics.
Research or practical experience in advanced approximation and compression techniques, including noise-aware or physics-constrained training.
Translation Company Name: Unconventional, Inc.