Sr. Staff, ML Researcher - LLM Algorithmic Optimization
d-MatrixJob Title
Sr. Staff, ML Researcher - LLM Algorithmic Optimization
Role Summary
Invent, design, and implement efficient algorithms to optimize large language model (LLM) inference on DNN accelerators. Work on the Algo team with mathematicians, ML researchers, and engineers to translate algorithmic and numerical techniques into high-impact, hardware-aware solutions.
Experience Level
Senior — 5+ years of hands-on experience in machine learning research, algorithm development, or related technical roles.
Responsibilities
Primary responsibilities include research, prototyping, and production-focused implementation of algorithms that improve LLM inference performance on custom accelerators.
- Research and develop algorithms to reduce latency, memory, and compute for LLM inference on DNN accelerators.
- Design and implement prototypes and production code, primarily in Python with object-oriented practices.
- Analyze numerical properties, stability, and error behavior of proposed methods.
- Profile and benchmark algorithms on target hardware; identify bottlenecks and optimize implementations.
- Collaborate closely with hardware architects, ML engineers, and software teams to map algorithms to accelerator constraints.
- Document designs, present results to stakeholders, and drive adoption of successful techniques.
- Mentor colleagues and contribute to the team’s technical roadmap.
Requirements
Core technical and experience requirements. Degree specifics are listed under Education Requirements below.
- 5+ years of hands-on experience in ML research, algorithm development, or closely related applied research roles.
- Strong mathematical and analytical skills relevant to numerical methods and ML.
- Proficiency in Python and object-oriented software design.
- Experience implementing numerical algorithms and optimizing performance (profiling, benchmarking).
- Able to collaborate across software and hardware teams and communicate technical results clearly.
- Nice-to-have: experience with transformer architectures or large language models.
- Nice-to-have: experience mapping algorithms to accelerators, DNN runtimes, or low-level performance optimization.
Education Requirements
Posting specifies an MSc or PhD in mathematics, computer science, statistics, physics, or a related STEM field. The posting also associates this with an expectation of 5+ years of hands-on experience.
About the Company
Company: d-Matrix
Headquarters: Santa Clara, California, United States
d-Matrix is a Santa Clara–based startup developing highly programmable in-memory computing architectures and accompanying software to accelerate generative AI and other AI workloads, focusing on hardware-software co-design for cloud and edge applications.
