Full-Time

Senior Software Engineer

ML Compiler, Frameworks and Performance

Updated on 9/3/2026

Google

Google

10,001+ employees

Tech giant offering search, ads, cloud

No salary listed

Bengaluru, Karnataka, India

In Person

Bachelor's

Category
Software Engineering (1)
Required Skills
Graphics Processing Unit (GPU)
Python
TensorFlow
Keras
PyTorch
Machine Learning
C/C++

Get referred to Google

See people who can refer or advise you

Requirements
  • A bachelor's degree or equivalent practical experience is required.
  • At least 5 years of software development experience in C++ or Python is required.
  • Experience in machine learning infrastructure development or machine learning performance engineering is required.
Responsibilities
  • Drive continuous improvements to the machine learning software and hardware stack for Google first-party teams and Google Cloud Platform customers.
  • Develop an understanding of Google's machine learning training and serving stack, including JAX, PyTorch, Accelerated Linear Algebra, and the runtime stack.
  • Identify opportunities to improve machine learning workload efficiency through performance debugging of workloads and custom kernels, and build solutions to deliver those improvements.
  • Partner with teams responsible for different parts of the machine learning stack to understand performance optimization use cases.
  • Work with open-source machine learning inference frameworks such as TorchTPU, vLLM, and SGLang to identify performance bottlenecks.
  • Support new machine learning paradigms, including horizontal scaling for upcoming TPU chips, by contributing across the stack and to performance analysis tools.
Desired Qualifications
  • Experience with machine learning compilers and their internals, including writing compiler optimization passes.
  • Experience with accelerator hardware architectures such as TPUs and GPUs.
  • Experience with machine learning inference frameworks such as vLLM, SGLang, and Pathways.
  • Experience with machine learning frameworks such as TensorFlow, JAX, PyTorch, and Keras.

Google provides internet services and products, including a widely used search engine and a large ecosystem of tools for individuals and businesses. It works by processing user queries and ranking results with relevance and personalized data, while its advertising platforms show targeted ads to users. Its products span cloud computing (Google Cloud), productivity software (Google Workspace), and hardware like Pixel phones and Nest devices, all connected through data and machine learning. Its goal is to organize the world’s information so it is universally accessible and useful, while supporting a sustainable business model built on ads and a growing software, cloud, and hardware portfolio.

Company Size

10,001+

Company Stage

IPO

Headquarters

Mountain View, California

Founded

1998

Get referred to Google

See people who can refer or advise you

Simplify Jobs

Simplify's Take

What believers are saying

  • Google Cloud revenue rose 82% to $24.8 billion in Q2 2026.
  • Search queries hit an all-time high last quarter despite AI feature expansion.
  • Google hired Windsurf CEO Varun Mohan and staff, strengthening AI coding tools.

What critics are saying

  • EU publishers filed an antitrust complaint on February 10, 2026 over AI Overviews.
  • Britain forced Google to let publishers opt out of AI search on June 3, 2026.
  • AI Mode threatens ad economics and search traffic as users bypass blue links.

What makes Google unique

  • Google Search AI Mode surpassed 1 billion monthly users in May 2026.
  • Gemini 3.5 Flash now powers Search, AI Studio, Android Studio, and AI Mode.
  • Google Cloud combines infrastructure, TPU sales, and Gemini distribution across enterprise contracts.

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

Benefits

Hybrid Work Options

Remote Work Options

Performance Bonus

Company Equity

Company News

LinkedIn
Jul 14th, 2025
Is has just been announced Google.org provided a €1,000,000 grant to electricityMap, a leading provider of granular data on electricity and carbon intensity. | Olivier Corradi

Is has just been announced Google.org provided a €1,000,000 grant to electricityMap, a leading provider of granular data on electricity and carbon intensity. In addition to bringing new data to the platform and making electricity data more accessible, electricityMap will use the funding to help policy makers, academic researchers and the private sector understand the key factors of sustainable electricity consumption, and drive demand for carbon-free solutions. Check out our open positions here: https://lnkd.in/eAyzmDCP | 110 comments on LinkedIn

Business Insider
Jul 14th, 2025
Business Insider - Latest News in Tech, Markets, Economy & Innovation

Business Insider tells the global tech, finance, stock market, media, economy, lifestyle, real estate, AI and innovative stories you want to know.

Dealroom.co
Jul 14th, 2025
Windsurf company information, funding & investors

Windsurf, ai coding assistant. founded as exafunction, a gpu virtualization platform, pivoted to codeium, now called windsurf. Here you'll find information about their funding, investors and team.

The Information
Jul 13th, 2025
Google to Pay $2.4 Billion for Windsurf Staff, IP After Startup Ends OpenAI Talks

OpenAI’s discussions to buy Windsurf, the maker of a popular artificial intelligence coding assistant formerly known as Codeium, have ended. Instead, Google will hire Windsurf’s CEO, Varun Mohan, and some of the startup’s staff, both Google and Windsurf said. The monthslong talks between OpenAI ...

PR Newswire
Jun 26th, 2025
Ubie Health Secures $125M for Expansion

Ubie announced its cumulative funding has surpassed $125 million, with recent investments from companies like Google and JAPAN POST CAPITAL. The US Symptom Checker platform has grown 85% organically, reaching 4 million users. Ubie is expanding into generative AI and provider solutions, leveraging Mayo Clinic's Platform Accelerate program. The company aims to enhance patient education and refine its platform using de-identified data from 3 million Mayo Clinic patients.