N

NVIDIA

Designs the chips and software that power AI

Performance Software Engineer Intern - Deep Learning Libraries

Internship
No salary listed
Master's, PhD
Beijing, China+1 moreMore locations: Shanghai, China
In Person
Software EngineeringSystems Software Engineering

About the job

Requirements
  • Pursuing a Master's or PhD degree in Computer Science, Computer Engineering, Applied Mathematics, or a related field.
  • Demonstrated strong programming and software design skills, including debugging, performance analysis, and test design.
  • Experience with performance-oriented parallel programming, including non-GPU programming such as OpenMP or pthreads.
  • A solid understanding of computer architecture and some experience with assembly programming.
Responsibilities
  • Write highly tuned compute kernels to perform core deep learning operations, including matrix multiplies, mixture-of-experts, and attention.
  • Follow software engineering best practices, including supporting regression testing and continuous integration and continuous delivery flows.
  • Collaborate with the compiler team on generating optimal assembly code.
  • Collaborate with deep learning training and inference performance teams to determine which layers require optimization.
  • Collaborate with hardware and architecture teams on the programming model for new deep learning hardware features.
  • Identify bottlenecks, optimize resource utilization, and improve throughput.
Desired Qualifications
  • Tune deep learning library kernel code.
  • Use CUDA for GPU programming.
  • Apply numerical methods and linear algebra.
  • Use LLVM, TVM tensor expressions, or TensorFlow MLIR.

About the company

N

NVIDIA

View
cbWebsite

NVIDIA designs the chips and software that power artificial intelligence and sells complete data center systems to cloud providers, AI labs, and other businesses, along with GeForce graphics cards for gamers. Its GPUs, chips that perform thousands of calculations at once, became the main hardware for AI after researchers used them to train the AlexNet image-recognition model in 2012. NVIDIA connects many GPUs with its own processors and networking gear into servers and rack-sized computers, and developers program them through its CUDA software. These systems bring in most of its sales, and it also makes chips for cars and robots. Jensen Huang co-founded NVIDIA in 1993, and it is based in Santa Clara, California.

Company Size

10,001+

Company Stage

IPO

Headquarters

Santa Clara, California

Founded

1993

Get referred to NVIDIA

See people who can refer or advise you

Simplify Jobs

Simplify's Take

What believers are saying

  • AI demand is still surging: data center sales rose 117% to $89 billion in the quarter ending July 2026.
  • Stock grants are worth more: shares hit a record in October 2026, up 28% for the year.
  • Engineering dominates hiring: 1,722 of NVIDIA's 2,686 job listings in October 2026 were in engineering.

What critics are saying

  • Big buyers are hedging: Meta's February 2026 agreement covers 6 gigawatts of AMD chips.
  • Financing its buyers is getting harder: lenders pushed back on NVIDIA's $500 billion chip-backed loan plan in October 2026.
  • Staff described long hours: Bloomberg's August 2024 interviews reported seven-day weeks, late nights and shouting matches.

What makes NVIDIA unique

  • Rivals must match its software too: over 7.5 million developers use CUDA and other NVIDIA tools to program its chips.
  • The CEO has a broad direct staff: Jensen Huang says over 60 people report to him, without one-on-ones.
  • People rarely leave: NVIDIA's yearly staff turnover is 3.7%, versus a 21.9% average across the tech industry.

Help us improve and share your feedback! Did you find this helpful?

Benefits

  • Company Equity
  • 401(k) Company Match

Growth & Insights and Company News

6 month growth

1%

1 year growth

0%

2 year growth

-1%
Bitcoin Insider
Oct 27th, 2026
Nvidia's $3B Meta stake fuels circular AI economy fears as Meta seeks $95B

Nvidia is buying $3 billion of shares in Meta, one of its biggest customers, as Meta moves to raise up to $95 billion in equity to fund its AI spending. Nvidia, AMD and Broadcom have agreed to buy as much as $5 billion of the shares between them. Nigel Green, chief executive officer of deVere Group, called it a sign of a "circular AI economy" in which chipmakers buy shares in the companies that buy their chips. "Sooner or later, someone outside the circle has to pick up the bill," he said. Meta will start with $35 billion of underwritten offerings, followed by a $60 billion at-the-market sale from early 2027. Its projected 2026 capital expenditure is $135 billion to $145 billion. Nvidia put $100 billion into OpenAI in 2025 and has joined 121 funding rounds this year, up from 13 in 2022.

Yahoo Finance
Oct 27th, 2026
Nvidia investors should watch big tech capex as Amazon, Microsoft, Alphabet report

Microsoft, Alphabet, and Amazon report earnings this week, and Nvidia investors should watch one number: capital expenditure. Microsoft and Alphabet report on October 28, and Amazon follows on October 29, all after the close of trading. The three are Nvidia customers, so higher spending on data centres means more demand for its chips. Microsoft guided to $175 billion for fiscal 2027. Alphabet expects $195 billion to $205 billion for 2026, up from $180 billion to $190 billion. Amazon raised its forecast by $20 billion to $220 billion. All three cite demand that outpaces capacity. Nvidia's revenue rose 106% to $96 billion last quarter, with more than 90% from data centres. Any slowdown in big tech spending would hurt Nvidia, at least in the short term.

TechCrunch
Oct 27th, 2026
Nvidia backs Indian startup Simplismart in $14M round to run AI models on fewer GPUs

Nvidia has backed Indian startup Simplismart in a $14 million Series B round led by Dallas Venture Capital. Accel, Shastra VC, and Micromax also took part. The deal values the four-year-old company at about $100 million. Simplismart's software sits between AI models and computing infrastructure, tuning how models run so companies can get more from their existing GPUs. It says it helped video platform InVideo cut GPU serving costs by 56%. The Bengaluru-based startup's annualised revenue run rate has grown 14-fold in six months to an eight-figure sum. More than half of its revenue comes from the US, where it plans to expand and grow its team of 58 people.

Yahoo Finance
Oct 26th, 2026
Nvidia shares rise 1.8% to $214.78 as Cadence launches AI models built on its tech

Nvidia shares rose 1.81% to $214.78 at 10:30 ET on Monday after Cadence introduced AI models built partly on Nvidia's Nemotron and OpenShell technologies. The move gives Nvidia another route into chipmakers' engineering workflows, beyond the processors that power AI systems. Cadence's Maestro can run on customers' private infrastructure, Cadence's cloud, or other public clouds. Early access is scheduled for the fourth quarter, according to Reuters. Nvidia's latest quarterly revenue reached $96.2 billion. Data centres contributed $89.0 billion, roughly 92.5% of sales. The commercial test comes next. Investors want evidence of paying customers, sustained inference demand, and licensing or infrastructure sales tied to the installations.

Yahoo Finance
Oct 26th, 2026
Nvidia reportedly sold out through 2028 as analysts forecast $692B revenue

Nvidia is sold out through 2028, according to Palo Alto Networks chief executive Nikesh Arora. On the All-In podcast, he said building his own AI data centres would be "kind of an empty threat," partly because "you talk to Jensen, he can't give you chips for 2028." AI critic Ed Zitron says Nvidia is creating the shortage to push customers to buy now. Yet Wedbush's Dan Ives put demand to supply at "13 to 1," and renting an H100 chip costs 45% more than in December, according to Social Capital. Analysts expect revenue to rise 68% to $692 billion in the year ending January 2028, in line with management's guidance. Nvidia trades at 17x forward earnings, its lowest multiple in three years. Microsoft, Alphabet, and Amazon report capital spending plans next week.