As part of the Safeguards team, you'll design and implement ML infrastructure that powers Claude safety. Your work will directly contribute to making AI systems more trustworthy and aligned with human values, ensuring our models operate safely as they become more capable.

Responsibilities:

• Design and build scalable ML infrastructure to support real-time and batch classifier and safety evaluations across our model ecosystem
• Build monitoring and observability tools to track model performance, data quality, and system health for safety-critical applications
• Collaborate with research teams to productionize safety research, translating experimental safety techniques into robust, scalable systems
• Optimize inference latency and throughput for real-time safety evaluations while maintaining high reliability standards
• Implement automated testing, deployment, and rollback systems for ML models in production safety applications
• Partner with Safeguards, Security, and Alignment teams to understand requirements and deliver infrastructure that meets safety and production needs
• Contribute to the development of internal tools and frameworks that accelerate safety research and deployment

You may be a good fit if you:

• Have 5+ years of experience building production ML infrastructure, ideally in safety-critical domains like fraud detection, content moderation, or risk assessment
• Are proficient in Python and have experience with ML frameworks like PyTorch, TensorFlow, or JAX
• Have hands-on experience with cloud platforms (AWS, GCP) and container orchestration (Kubernetes)
• Understand distributed systems principles and have built systems that handle high-throughput, low-latency workloads
• Have experience with data engineering tools and building robust data pipelines (e.g., Spark, Airflow, streaming systems)
• Are results-oriented, with a bias towards reliability and impact in safety-critical systems
• Enjoy collaborating with researchers and translating cutting-edge research into production systems
• Care deeply about AI safety and the societal impacts of your work

Strong candidates may have experience with:

• Working with large language models and modern transformer architectures
• Implementing A/B testing frameworks and experimentation infrastructure for ML systems
• Developing monitoring and alerting systems for ML model performance and data drift
• Building automated labeling systems and human-in-the-loop workflows
• Experience in trust & safety, fraud prevention, or content moderation domains
• Knowledge of privacy-preserving ML techniques and compliance requirements
• Contributing to open-source ML infrastructure projects

We encourage you to apply even if you do not believe you meet every single qualification. Not all strong candidates will meet every single qualification as listed.

XML job scraping automation by YubHub

Location

San Francisco

Employment Type

Intern

Location Type

On-site

Department

Applied AI

Deadline to Apply

March 11, 2026 at 3:00 AM EDT

Compensation

• $67 per hour

About the Team

The Applied team works across research, engineering, product, and design to bring OpenAI’s technology to the world. We seek to learn from deployment and broadly distribute the benefits of AI, while ensuring that this powerful tool is used responsibly and safely. We aim to make our innovative tools globally accessible, transcending geographic, economic, or platform barriers. Our commitment is to facilitate the use of AI to enhance lives, fostered by rigorous insights into how people use our products.

About the Role

A systems research internship is for people who love the real-world intersection of systems-engineering and research: you’ll investigate a hard systems problem, build something meaningful, and measure it carefully. The goal is practical impact—making Applied Systems better: more efficient, more scalable, and more reliable.

OpenAI is currently recruiting for candidates interested in a 13-week, paid, in-person internship based in our San Francisco office during Summer 2026. In some cases, it may be extended for an additional 13 weeks (for a total of up to 26 weeks), based on team needs, candidate interest, and performance.

In this role, you will typically focus on improving real systems in areas like:

• Distributed systems & storage (throughput, latency, consistency, durability)

• Compute & scheduling (GPU/accelerator utilization, job orchestration, queuing)

• Performance engineering (profiling, bottlenecks, scalability, capacity planning)

• Reliability & observability (fault tolerance, monitoring, incident learning)

• Networking & data pipelines (data movement, caching, streaming efficiency)

• Systems for ML (training/inference performance, evaluation infrastructure, tooling)

Most projects involve some of these steps:

• Defining a clear hypothesis (“we think X will reduce latency by Y under Z”)

• Instrumenting existing production systems, gathering metrics and detailed analysis to validate the hypothesis

• Building or modifying a real system (prototype or production-quality improvements when appropriate)

• Running experiments/benchmarks and analyzing results

• Communicating tradeoffs and recommendations clearly

• Publishing the research work in technical journals and conferences

Your background looks something like:

• Currently pursuing a PhD in Computer Science, Computer Engineering, or relevant technical field

• Proficiency with Coding in c++, Java, python, rust, etc

• Doing ongoing research on systems topics such as DL/ML, information retrieval, systems security and cryptography, databases, networking, distributed systems, and compilers, etc

• Ability to move fast in an environment where things are sometimes loosely defined and may have competing priorities or deadlines

About OpenAI

OpenAI is an AI research and deployment company dedicated to ensuring that general-purpose artificial intelligence benefits all of humanity. We push the boundaries of the capabilities of AI systems and seek to safely deploy them to the world through our products. AI is an extremely powerful tool that must be created with safety and human needs at its core, and to achieve our mission, we must encompass and value the many different perspectives, voices, and experiences that form the full spectrum of humanity.

XML job scraping automation by YubHub