Machine Learning Engineer (Model Quantization)

Location: Austin, TX or Sunnyvale, CA; full-time onsite position

Salary: $160K–$215K annually, plus equity

Company

hirify.global is an AI hardware startup developing silicon photonics and programmable metasurface-based optical inference engines for highly efficient, high-throughput AI computation.

What you will do

• Develop hardware-aware post-training quantization methods for LLMs, diffusion models, and other machine learning applications.
• Investigate non-convex, discrete, constrained, and second-order optimization approaches for quantization and preconditioning.
• Design controlled numerical experiments and build research-quality, reproducible implementation and evaluation harnesses.
• Adapt open-source and customer models across PyTorch, Triton, JAX, and other frameworks.
• Develop re-quantization, retraining, and model adaptation techniques that minimize accuracy loss during precision reduction.
• Collaborate with hardware, software, and architecture teams to optimize GEMM operations and model architectures for optical compute.

Requirements

• PhD or equivalent research experience in machine learning, applied mathematics, optimization, numerical analysis, computer science, or a related field.
• 5+ years of machine learning engineering experience, including at least 3 years focused on model optimization and deployment.
• Research or advanced engineering experience in neural network quantization, model compression, numerical optimization, or efficient inference.
• Strong knowledge of numerical linear algebra and experience with advanced optimization methods.
• Strong proficiency in PyTorch and familiarity with JAX, Triton, and TensorFlow.
• Hands-on experience with transformer architectures, LLMs, diffusion models, controlled numerical experiments, and research collaboration.

Nice to have

• Experience with INT8, FP8, or lower-precision inference optimization.
• Background in analog or optical computing, in-memory computing, or matrix-vector multiplication acceleration.
• Knowledge of randomized numerical linear algebra, sketching, structured transforms, vector quantization, lattice methods, learned codebooks, or rate-distortion techniques.
• Publications in quantization, optimization, numerical linear algebra, model compression, or efficient machine learning.
• Experience with large-scale batch inference and LLM prefill versus decode optimization.

Culture & Benefits

• Health plan premiums covered at 100% for employees and dependents, with HSA contributions.
• Unlimited paid time off.
• 401(k) matching and stock option opportunities.
• Dental, vision, life, hospital, critical illness, and accident insurance options.
• Flexible benefits selection with cash-back options for unused plans.