onsite
Applied Scientist, Machine Learning Accelerator - Amazon
ML Engineer
Design and implement pre‑silicon verification frameworks for a custom Machine Learning accelerator, using formal methods, SMT solving, and software tooling to generate realistic stimulus and assembly tests.
About the role
Key Responsibilities
- Develop and maintain a verification framework that generates legal, realistic stimulus for a custom ML accelerator chip.
- Apply formal methods and SMT solvers to model design behavior and solve constraint problems for test generation.
- Write high‑performance software components (Python, C++) to automate assembly‑level test creation and execution.
- Collaborate with hardware designers to integrate verification flows into the pre‑silicon design cycle.
- Analyze verification results, debug failures, and iterate on models to improve coverage and accuracy.
Requirements
- Strong programming skills in Python and C++ with experience building verification or testing tools.
- Hands‑on experience with formal verification techniques and SMT solvers (e.g., Z3, Boolector).
- Background in hardware design or ASIC/FPGA verification, preferably for ML accelerators.
- Solid understanding of machine learning concepts and how they map to custom silicon architectures.
- Excellent problem‑solving abilities and ability to work cross‑functionally with hardware and software teams.
Skills
pythoncmachine learning