Tesla

Makes electric cars, robotaxis and grid batteries

NPU Performance Engineer - SI Hardware

Hardware Validation & TestingConfirmed live in the last 24 hours

Full-timeAustin, TX, USANo salary listedIn-person

About the Job

Requirements

A degree in Engineering, Electrical Engineering, Computer Science, or a related field, or equivalent experience.
At least 1 year of industry experience in performance verification of NPUs, SI accelerators, or ML hardware.
A strong track record writing, optimizing, and verifying high-performance kernels for NPUs or similar tensor-core architectures, including direct experience supplying kernels to ML compiler teams.
Deep expertise correlating ML model behavior, kernel execution, compiler integration, and NPU hardware across the verification stack, including RTL simulation, emulation, and post-silicon.
Mastery of Python and C++ for verification tools, automation frameworks, kernel development, and analysis pipelines.
A solid understanding of NPU microarchitecture, memory hierarchies, dataflow scheduling, on-chip interconnects, parallel computing principles, and ML compiler flows.

Responsibilities

Own end-to-end performance verification of custom NPUs, ensuring real ML workloads and handwritten kernels meet power, latency, utilization, and throughput targets.
Write, optimize, and verify high-performance kernels for critical ML operators, with many kernels integrated and used by the ML compiler team.
Develop verification suites and correlation methodologies that map high-level ML models through custom kernels to NPU hardware behavior.
Collaborate with ML compiler engineers and hardware architects to translate model requirements into optimized kernels and validated hardware performance.
Debug performance discrepancies across simulation, emulation, FPGA, and post-silicon environments, and root-cause issues involving models, kernels, compilers, and NPU microarchitecture.
Build automated regression frameworks, performance dashboards, and correlation tools to accelerate verification cycles and provide visibility into NPU health.
Analyze utilization, bandwidth, latency, and power metrics, and establish correlation between pre-silicon predictions and real silicon measurements.
Document and present verification findings, correlation results, and optimization breakthroughs to influence NPU roadmaps and silicon strategy.

Desired Qualifications

Kernel-writing and ML compiler experience is a plus.

Category	Tesla SI
Location	Austin, Texas
Requisition ID	267771
Type	Full-time

Job Description

The Tesla SI Hardware team is at the forefront of revolutionizing artificial intelligence through cutting-edge hardware innovation. Comprising brilliant engineers and visionaries, the team designs and develops advanced SI inference chips tailored to accelerate Tesla’s machine learning capabilities. A key part of this effort is Dojo, Tesla's custom supercomputer system built to efficiently train massive neural networks on vast video data from the fleet. The work of Tesla's SI Hardware team powers the neural networks behind Full Self-Driving (FSD), and Tesla humanoid robot, Optimus, pushing the boundaries of computational efficiency and performance. By creating custom silicon and optimized architectures, the team ensures Tesla remains a leader in SI-driven automotive and energy solutions, shaping a future where intelligent machines enhance human life.<br><br>Join Tesla’s elite SI Hardware team and become a driving force behind the next generation of autonomous intelligence. In this high-impact role, you will own performance verification of next-generation NPUs, stress-testing and correlating real ML models and handwritten kernels all the way down to silicon behavior. This is hands-on, deep-dive work that blends kernel development, compiler integration, and rigorous verification to ensure our NPUs deliver groundbreaking efficiency, throughput, and determinism. You’ll write kernels destined for the ML compiler team while relentlessly validating performance across the full stack — helping power Tesla’s Full Self-Driving and Optimus platforms at the absolute edge of what’s possible.

Job Responsibilities

Own end-to-end performance verification of Tesla’s custom NPUs, ensuring real ML workloads and handwritten kernels meet aggressive power, latency, utilization, and throughput targets
Write, optimize, and verify high-performance kernels for critical ML operators, with many kernels directly integrated and consumed by the ML compiler team
Develop advanced verification suites and correlation methodologies that map high-level ML models through custom kernels down to actual NPU hardware behavior
Collaborate closely with ML compiler engineers, and hardware architects to translate model requirements into optimized kernels and validated hardware performance
Debug complex performance discrepancies across simulation, emulation, FPGA, and post-silicon environments — root-causing issues at the intersection of model, kernel, compiler, and NPU microarchitecture
Build automated regression frameworks, performance dashboards, and correlation tools that accelerate verification cycles and provide real-time visibility into NPU health
Perform deep analysis on key metrics (utilization, bandwidth, latency, power) and establish iron-clad correlation between pre-silicon predictions and real silicon measurements
Document and present verification findings, correlation results, and optimization breakthroughs to influence future NPU roadmaps and Tesla’s SI silicon strategy

Job Requirements

Degree in Engineering, Electrical Engineering, Computer Science, or related field or equivalent experience
1+ years of intense industry experience in performance verification of NPUs, SI accelerators, or ML hardware
Strong track record writing, optimizing, and verifying high-performance kernels targeted at NPUs or similar tensor-core architectures, with direct experience supplying kernels to ML compiler teams
Deep expertise correlating ML model behavior → kernel execution → compiler integration → NPU hardware across the full verification stack (RTL simulation, emulation, and post-silicon)
Mastery of Python and C++ for building high-performance verification tools, automation frameworks, kernel development, and analysis pipelines
Solid understanding of NPU microarchitecture, memory hierarchies, dataflow scheduling, on-chip interconnects, parallel computing principles, and ML compiler flows
Kernel writing and ML compiler experience is a plus
Exceptional debugging skills with a proven ability to resolve the toughest NPU performance bottlenecks under tight deadlines
Thrives in a fast-paced, ownership-driven, small-team environment with outstanding technical communication skills

About the Company

Tesla

Makes electric cars, robotaxis and grid batteries

Tesla makes electric cars and batteries that store power for homes and utilities, and it develops self-driving software and humanoid robots. In 2023, its Model Y became the first electric vehicle to rank as the world's best-selling car. Buyers order the Model 3 sedan, Model Y SUV, or Cybertruck pickup directly from Tesla instead of through dealers, and they charge on its Supercharger fast-charging network. It also sells Megapack and Powerwall batteries, runs a robotaxi service, and charges drivers a monthly fee for self-driving software that still needs their supervision. Tesla was founded in 2003, with Elon Musk as its lead early investor and a recognized co-founder, and it is based in Austin, Texas.

Company size: 10,001+
Company stage: IPO
Headquarters: Austin, TX
Founded: 2003

Recently Posted Jobs

Explore more opportunities at Tesla

See more jobs