Algorithm Engineer, Co-Design for Analog AI Hardware

Normal Computing Corporation; New York (NY); On-site; USD 140,000 - 210,000; Full time

Job summary

Normal Computing seeks researchers to develop algorithms for transformer inference on stochastic analog processing-with-memory hardware. You will co-design with hardware teams to map workloads to the chip's native dynamics and validate approaches on silicon or high-fidelity simulators.

The role emphasizes hands-on implementation, real hardware experience, and collaboration across software, architecture, and hardware disciplines to push the limits of energy-efficient AI inference.

Qualifications

• Deep understanding of large model inference: attention mechanisms, KV cache, long-context decoding, and memory bandwidth.
• Experience with inference optimization: quantization, sparsity, kernel fusion, or memory-efficient attention.
• Familiarity with stochastic systems, probabilistic methods, numerical analysis, or analog computation.
• Experience implementing algorithms close to hardware, not just in high-level frameworks.
• Comfort reasoning from first principles about what a novel substrate can do efficiently.
• Track record of taking ideas from theory to working implementation on real hardware.
• Strong programming skills in Python and at least one systems language.
• Collaborative instinct and ability to work across hardware, architecture, and software teams.

Responsibilities

• Algorithm Development for transformer inference workloads on stochastic analog processing-with-memory hardware.
• Software/Hardware Co-Design with hardware and architecture teams to shape native compute.
• Numerical Methods exploiting thermal noise and analog dynamics.
• Evaluation & Benchmarks on real hardware or high-fidelity simulation.
• Workload Translation: map model workloads into hardware constraints and opportunities.
• Rapid Prototyping as hardware evolves from simulation to silicon.

Skills

Large model inference

Attention mechanisms

KV cache

Long-context decoding

Memory bandwidth

Inference optimization

Python

C/C++

Systems language

Hardware-software co-design

Tools

Python

C/C++

CUDA

Job description

Normal Computing seeks researchers to develop algorithms for transformer inference on stochastic analog processing-with-memory hardware. You will co-design with hardware teams to map workloads to the chip's native dynamics and validate approaches on silicon or high-fidelity simulators.

The role emphasizes hands-on implementation, real hardware experience, and collaboration across software, architecture, and hardware disciplines to push the limits of energy-efficient AI inference.