Role Overview

The Software Engineer, Triton Compiler position at OpenAI represents a rare opportunity to work at the intersection of compilers, kernels, and cutting-edge AI hardware. Based in San Francisco, California, this role sits within the performance-critical software stack that powers OpenAI's custom silicon accelerators.

OpenAI's mission to discover and enact the path to safe, beneficial AGI requires technical breakthroughs across generative modeling, reinforcement learning, large-scale optimization, and active learning. The Compiler/Kernels team delivers these breakthroughs by designing compilers, languages, and high-performance kernels that enable researchers to fully exploit first-party accelerators.

In this role, you'll advance Triton and its backend, develop novel compiler passes, and create tooling for deeply optimized kernels targeting brand-new hardware. You'll partner directly with the hardware team to unlock capabilities supporting next-generation frontier models. If you thrive in ambiguous technical spaces and love shaping how AI hardware gets programmed, this is your chance to make history.

Software Engineers on this team demonstrate strong fundamentals in designing ML systems, writing reliable low-level code, and pushing performance boundaries. Every project shares the same goal: advance the AI capability frontier.

Key Responsibilities

As a Software Engineer, Triton Compiler, your day-to-day will include:

• Designing and building compilers, domain-specific languages, and high-performance kernels for OpenAI's custom silicon
• Advancing the Triton compiler ecosystem and its backend infrastructure
• Developing innovative compiler passes that optimize AI workloads for new hardware architectures
• Creating developer tooling that enables writing fast, correct, deeply-optimized kernels
• Partnering closely with hardware engineers to co-design software-hardware interfaces
• Optimizing core ML systems to achieve previously impossible performance levels
• Writing highly reliable low-level C++, CUDA, and assembly code for performance-critical paths
• Profiling, debugging, and tuning kernels across diverse accelerator architectures
• Integrating compiler technologies with PyTorch, JAX, and other ML frameworks
• Identifying highest-leverage performance bottlenecks across the AI stack
• Contributing to open-source Triton advancements and MLIR/LLVM ecosystem
• Shaping programming models for entirely new classes of AI hardware
• Ensuring kernel correctness through rigorous testing and verification
• Collaborating with researchers to translate novel algorithms into production kernels

These responsibilities demand engineers who own problems end-to-end, rapidly learn new domains, and deliver results that move the AGI needle.

Qualifications

Successful candidates will demonstrate:

• 3+ years of relevant experience in systems programming, compilers, ML frameworks, or performance engineering
• Proven ability to own complex technical problems from inception to production
• Experience rapidly learning new hardware/software domains as needed
• Strong fundamentals in ML systems design and low-level optimization
• Track record writing production-quality low-level code (C++, CUDA, assembly)
• Bonus: Open-source contributions to PyTorch, JAX, TensorFlow, LLVM, MLIR, or XLA
• Bonus: Experience programming custom accelerators, GPUs, or TPUs
• Self-directed mindset thriving in ambiguous, rapidly-evolving technical environments
• Excitement for identifying and solving highest-leverage problems
• Deep systems knowledge spanning compilers, kernels, runtime systems, and hardware

OpenAI values diverse engineering backgrounds but prioritizes demonstrated impact over pedigree.

Salary & Benefits

Competitive compensation for Software Engineer, Triton Compiler roles in San Francisco ranges from $250,000 - $450,000 base salary, plus equity, bonus, and comprehensive benefits. Total compensation reflects experience level and impact.

OpenAI offers industry-leading benefits including:

• Comprehensive health, dental, vision coverage
• Mental health support and therapy benefits
• 401(k) with generous company match
• Unlimited PTO policy
• Paid parental leave for primary/secondary caregivers
• Fertility benefits including IVF coverage
• Learning & development stipend
• Daily healthy meals and snacks
• Gym membership reimbursement
• Commuter benefits
• Relocation support
• Equity in a mission-driven company reshaping humanity's future

Why Join OpenAI?

OpenAI isn't just another AI company—it's the organization pushing AGI boundaries while prioritizing safety and human benefit. Your work on Triton Compiler directly enables frontier models that could transform society.

Join a team of exceptional engineers solving problems no one else can. Work with custom silicon designed from the ground up for AGI research. Shape programming models that will define AI hardware for decades.

OpenAI's culture emphasizes impact, ownership, and rapid learning. We hire self-directed builders excited by ambiguous challenges. You'll collaborate with world-class researchers, hardware experts, and systems engineers daily.

We're committed to diversity, equity, and inclusion—actively seeking perspectives that represent humanity's spectrum. Background checks comply with fair chance laws including San Francisco Fair Chance Ordinance.

How to Apply

Ready to build the software stack powering safe AGI? Submit your resume and a brief note explaining why you're excited about compilers, kernels, and OpenAI's mission. Highlight specific projects demonstrating end-to-end ownership and performance impact.

We're hiring rapidly—strong candidates hear back within 1 week. OpenAI is an equal opportunity employer valuing diverse experiences and backgrounds.

Keywords: Triton compiler jobs San Francisco, OpenAI software engineer compiler, AI hardware programming careers, ML kernel optimization roles, custom silicon compiler engineer