AI Jobs in San Francisco

View detail

Translate business requirements into requirements for AI/ML models; prepare data to train and evaluate AI/ML/DL models; build AI/ML/DL models by applying state-of-the-art algorithms, especially transformers; leverage existing algorithms from academic or industrial research when applicable; test, evaluate, and benchmark the AI/ML/DL models, and publish the models, data sets, and evaluations; deploy models in production by containerizing the models; work with customers and internal employees to refine the quality of the models; establish continuous learning pipelines for models with online learning or transfer learning; build and deploy containerized applications on cloud or on-premise environments.

View detail

Contributors design original computational statistics problems that simulate real mathematical research workflows, creating problems requiring Python programming to solve using libraries like Numpy, SciPy, and Sympy. They ensure problems are computationally intensive and require non-trivial reasoning chains in areas such as number theory, combinatorics, graph theory, and numerical analysis. The problems are based on real research challenges or practical mathematical applications. Contributors verify solutions using Python with standard mathematical libraries and document problem statements clearly with verified correct answers.

View detail

Contributors may generate prompts that challenge AI, evaluate AI-generated solutions for correctness, assumptions, and logic, improve AI reasoning to align with first principles and accepted standards, and apply structured scoring criteria to assess multi-step problem solving.

View detail

Contributors may generate prompts that challenge AI; evaluate AI-generated solutions for correctness, assumptions, and logic; improve AI reasoning to align with first principles and accepted standards; and apply structured scoring criteria to assess multi-step problem solving.

View detail

Design original computational statistics problems that simulate real mathematical research workflows; create problems requiring Python programming to solve using libraries such as Numpy, SciPy, and Sympy; ensure problems are computationally intensive and cannot be solved manually within reasonable timeframes; develop problems requiring non-trivial reasoning chains in areas like number theory, combinatorics, graph theory, and numerical analysis; base problems on real research challenges or practical applications from mathematical practice; verify solutions using Python with standard mathematical libraries; document problem statements clearly and provide verified correct answers.

View detail

Work on a small, high-caliber team building AI products for clients, from requirements gathering and prototyping through system design, development, testing, and deployment. Own features end-to-end and develop domain expertise across a range of AI use cases. Spend most of the time coding and frequently interact with clients to ensure the solutions meet their needs.

View detail

You will work on a small, high-caliber team building AI products for clients, setting technical direction, writing code, and serving as the go-to person when challenges arise. Spend approximately 75% of your time coding and 25% interacting with clients, including CTOs, to understand problems, evaluate tradeoffs, and ensure solutions meet their needs.

View detail

As a member of the Bots team, design, build, and scale systems that enhance user engagement with the AI-powered platform, including bot chat orchestration, AI image generation, AI video generation, and tooling for managing these features. Collaborate with cross-functional teams like product managers, designers, and data specialists to deliver high-quality, performant, and maintainable features. Experiment with and integrate new AI image, video, and voice generation technologies. Build tooling and infrastructure around various AI technologies. Gain exposure to the architecture and operations of a fast-growing social AI product. Contribute expertise to evolve team processes and technical infrastructure, ensuring scalability and reliability.

View detail

Design and implement worst-case demonstrations that concretely reveal AGI alignment risks for stakeholders, focusing on high-stakes use cases; develop adversarial and system-level evaluations based on these demonstrations and promote their adoption across OpenAI; create automated tools and infrastructure to scale automated red-teaming and stress testing; conduct research on failure modes of alignment techniques and propose improvements; publish influential internal or external papers that impact safety strategy or industry practice; collaborate with engineering, research, policy, and legal teams to integrate findings into product safeguards and governance; and mentor engineers and researchers to foster a culture of rigorous, impact-oriented safety work.

View detail

The AI Evaluation Engineer will analyze training and evaluation datasets to identify distributional gaps, labeling inconsistencies, and long-tail opportunities; design and execute labeling campaigns including the development of golden datasets and annotation guidelines; build and maintain dashboards to track model accuracy, regression trends, and product-specific KPIs; investigate failure modes through prompt clustering, error taxonomy development, and user intent classification; operationalize feedback loops by mining product telemetry and human-in-the-loop reviews for signal and translate these into data-driven model improvement strategies; partner with engineers and product managers to run structured A/B tests and human evaluations for new models or features; support the development of scalable data and evaluation infrastructure for LLMs and agents; and work with product, engineering, and legal teams to create clear and transparent processes for handling customer data in AI training, fine-tuning, and evaluation.