Top AI Specialist Jobs in San Francisco Bay Area, CA

Design realistic, specialty-specific clinical case scenarios and gold-standard attending-level responses, evaluate and grade AI-generated outputs using structured rubrics, provide detailed written feedback to improve model clinical reasoning, and participate in onboarding, calibration, and review sessions in a fully remote, asynchronous contractor role (~20 hours/week).

PhD-level physicists will design challenging graduate-level physics problems and contribute high-quality datasets to improve AI model reasoning, ensuring scientific rigor and consistency while collaborating remotely and asynchronously.

Lead design, validation, and continuous improvement of AI-assisted measurement and learning systems for K-12 assessments. Translate measurement intent into reproducible AI workflows, prototype AI-assisted content generation, ensure outputs meet validity, fairness, accessibility, and instructional standards, design validation and monitoring frameworks, partner with cross-functional teams (AI Science, Technology, Content), and contribute research, documentation, and thought leadership on intelligent assessment and learning systems.

The AI Operations Specialist will support CoinMarketCap's AI team by conducting feature evaluations, developing tools and automations, managing outreach, and collaborating with internal stakeholders to enhance AI operations.

Design, build, and deploy production AI systems and integrations (LLM-based) for law-firm clients. Create connectors, automation workflows, agents, and enablement programs, embed AI into client processes, and support internal AI adoption.

Sell coding-assistant and AI-native engineering solutions and co-sell NVIDIA developer-platform optimizations. Engage enterprise development leadership, design pilots and benchmarking services, scope consulting engagements, and collaborate within regional PODs to drive AI-accelerated application performance and developer productivity.

The AI/Machine Learning Specialist designs and implements AI solutions, collaborates with cross-functional teams, and continuously improves ML processes.

Design, build, and operate a secure, scalable GenAI platform on AWS (including Bedrock). Own platform lifecycle, enable LLM inference, RAG, Agents and guardrails, ensure observability, incident management, security, compliance, cost optimization, and reliability for production-grade regulated-financial environments. Provide Python-based development and support for microservices and REST APIs.

Lead complex, cross-functional AI programs end-to-end: translate strategy into integrated plans, drive multi-team execution, mitigate risks (data, model, compliance), apply ML lifecycle practices, use Agile/kanban methods, manage budgets/vendors, communicate to senior stakeholders, and mentor other program managers to improve AI program standards.

Own strategy, design, and optimization of Vanta's customer-facing chatbot and internal AI Copilot. Build and maintain AI-driven knowledge, design conversation flows in Intercom, analyze performance data, and collaborate cross-functionally to increase resolution rates and CSAT. Handle a small volume of support tickets to inform improvements and maintain product expertise.

Design, build, and optimize AI-driven customer success automations using no-code/low-code tools and LLM APIs. Identify use cases across the customer lifecycle, integrate solutions with enterprise platforms, apply prompt engineering and RAG patterns, monitor performance, and collaborate cross-functionally to scale AI-enabled workflows while maintaining responsible AI practices.

Review trainer-submitted image-question pairs and verify golden answers for accuracy and consistency. Flag and document incorrect, ambiguous, or incomplete responses, provide corrective feedback, track recurring errors, and escalate unclear cases. Ensure trainer output meets project quality standards.

Review trainers' completed annotation tasks comparing questions, images, and golden answers to verify correctness. Flag incorrect, ambiguous, or inconsistent responses, provide constructive feedback, track recurring errors, ensure outputs meet quality standards, and escalate unclear cases to the project team.