RESEARCHER, EFFICIENT INFERENCE

The role

You will research making models efficient through quantization, speculative decoding, sparse attention, distillation, and mixture-of-experts. The work involves algorithm design, evaluation, and partnering with engineers to implement methods in production.

What you will do

• Research and develop quantization methods
• Design and evaluate speculative decoding approaches
• Investigate training-time efficiency methods
• Run controlled experiments at production scale
• Co-design methods with the inference engineering team
• Read efficient ML and efficient inference literature
• Publish work and share findings internally
• Partner with model and training researchers

Required qualifications

• Strong track record of ML research on efficiency methods
• 5+ years of hands-on research experience
• Deep familiarity with both training and inference performance characteristics
• Fluent in PyTorch, Jax or equivalent
• Comfortable working at the kernel and serving-framework level
• Track record of moving efficiency research from prototype to production
• Strong statistical expertise
• Strong written communication
• Published research at NeurIPS, ICML, ICLR, MLSys, or comparable venues

Nice to have

• PhD in ML, systems, or related field
• Open-source contributions to quantization, speculative-decoding, or efficient-inference libraries
• Experience with hardware-aware optimization and accelerator-specific tooling
• Background in numerical methods, low-precision arithmetic, or approximate computation

About the company

makermaker.ai is a small, well-funded team building autonomous research agents for recursive self-improvement.