Source Job

• Own Agent retrieval accuracy and relevance.
• Drive automated resolution rates.
• Manage AI safety and trust.

Airtable is the no-code app platform that empowers people closest to the work to accelerate their most critical business processes. More than 500,000 organizations, including 80% of the Fortune 100, rely on Airtable to transform how work gets done.

• Investigate safety incidents, fraud cases, disputes, and policy violations.
• Gather evidence, document findings, and make recommendations for resolution.
• Partner with Operations, Legal, Support, and Risk teams on escalations.

Muvr is building the future of on-demand logistics and moving services. They connect customers with trusted drivers and crews to deliver large items quickly and reliably, ensuring safe interactions and a secure environment.

• Develop definitional protocols to identify trust-breaching customer experiences and recommend optimal response strategies.
• Establish processes for monitoring trust-breaching incidents and prepare periodic reports to leadership.
• Design Trust Policies to govern marketplace participation and develop internal training processes.

ezCater is a food for work technology company in the US, connecting workplaces to over 100,000 restaurants nationwide. They provide flexible and scalable solutions for employee meals and manage food spend, backed by 24/7 customer service. They are backed by top investors and values work/life harmony.

• Build new novel and long-horizon evaluations
• Develop novel measurement approaches for understanding how model capabilities emerge and evolve during RL training
• Lead strategic evaluation coverage across the company

Anthropic's mission is to create reliable, interpretable, and steerable AI systems, ensuring AI is safe and beneficial for users and society. They are a growing group of researchers, engineers, policy experts, and business leaders committed to building beneficial AI systems.

• Lead secure design reviews, threat modeling, and security-focused code reviews across the product and platform.
• Build and run Fieldguide’s vulnerability management program: scanning, triage, SLA-driven remediation tracking, and engineering coordination.
• Partner with Compliance to ensure technical controls satisfy framework requirements (SOC 2, ISO 27001, ISO 42001, FedRAMP).

Fieldguide is establishing a new state of trust for global commerce and capital markets through automating and streamlining the work of assurance and audit practitioners. They are based in San Francisco, CA, and built as a remote-first company that enables you to do your best work from anywhere.

• Investigate activity and disrupt abusive operations in partnership with our policy, legal, integrity, global affairs and security teams, including by conducting cross-internet and open source research
• Develop abuse signals and tracking strategies to help proactively detect harmful activity on our platform
• Communicate investigation findings from your work with stakeholders internally and, at times, externally

OpenAI's mission is to ensure that general-purpose artificial intelligence benefits all of humanity. They are an AI research and deployment company that pushes the boundaries of AI systems and seeks to safely deploy them to the world through their products.

• Own end-to-end delivery of ML-based safety features.
• Build and maintain ML models that safeguard AI-generated content.
• Design and implement RAG architectures to enhance detection capabilities.

Canva is a design platform that empowers everyone to create and share visual content. They have campuses in Sydney and Melbourne, and coworking spaces in other locations. They value passion, curiosity, and a willingness to learn.

• Follow company policies/standards to determine if activities or transactions are non-compliant/potentially suspicious and action as appropriate
• Ability to work / investigate alerts in a fast-paced environment while maintaining high quality standards
• Undertake ad project work and Financial Crime program enhancements

Affirm is reinventing credit to make it more honest and friendly, giving consumers the flexibility to buy now and pay later without any hidden fees or compounding interest. The company is building a financial crimes compliance program and seems to have a friendly and transparent culture.

• Design and execute complex jailbreak attempts to identify vulnerabilities in state-of-the-art models.
• Use your background in linguistics or social sciences to find "hidden" biases or harms that standard automated filters miss.
• Systematically rank LLM outputs to determine where safety guardrails are failing or succeeding.

We are building safer, more robust intelligence. We appear to be a small team with a culture that values asynchronous work and self-starters.

• Support regulatory and partner oversight to prepare and respond to information requests from regulators and partner banks.
• Automate compliance processes by partnering with internal teams and leveraging AI tools for efficiency.
• Support continuous monitoring using SQL, dashboards, and automation tools to detect potential compliance issues.

Tilt provides mobile-first financial products and machine learning-powered credit models. Valued as a next billion-dollar startup, it fosters a culture where every voice is valued and mutual respect is a priority.

• Lead the delivery and execution of the Behavioral Safety product roadmap
• Shape the vision and strategy for the future of Behavioral Safety
• Define, track and report on performance using relevant metrics

Reddit is a community-based platform that fosters open and authentic conversations. It has over 100,000 active communities and approximately 121 million daily unique visitors, making it one of the internet’s largest sources of information.

• Conduct regular data privacy and AI risk assessments and audits.
• Collaborate with cross-functional teams to develop and implement data privacy and AI policies.
• Monitor and analyze changes to data privacy and AI laws and regulations.

Docker makes app development easier so developers can focus on what matters. Their remote-first team spans the globe, united by a passion for innovation and great developer experiences; they’re growing fast and just getting started.

• Define and execute the strategy for end-to-end data lifecycle management.
• Lead the stitching of disparate data sources to create a high-fidelity, 360-degree view of a user’s debt history.
• Build and maintain the Feature Store and semantic layer required for possible AI features.

Summer is a Certified B Corp® committed to restoring financial freedom for Americans burdened by student debt. They combine policy expertise and technology to simplify college cost planning and loan repayment, partnering with employers, cities, and financial institutions.

• Architect the system and mentor the team, spend significant time hands-on in the codebase.
• Drive our strategy for SFT and RLHF/DPO; oversee the sourcing, labeling, and cleaning of diverse datasets.
• Design and train custom classifiers to detect and filter non-consensual or illegal content within an explicit environment.

EverAI is building the future of AI companionship, creating entirely new categories. They have 40 million users and are aiming for 100M and then 500M, consisting of an enthusiastic, passionate and hardworking team of 75 people.

• Design and curate evaluation datasets for retrieval quality.
• Measure retrieval quality using metrics like Recall@k, Precision@k, MRR, and NDCG@k.
• Conduct systematic error analysis on AI/ML system outputs; build structured failure taxonomies.

Jump empowers financial advisors, firms, and clients to thrive in the age of AI by automating tasks like meeting prep and compliance. As a Series A company, Jump has raised $30M and grown to 100+ employees including leaders from top companies and schools, fostering a culture of velocity, world-class standards, direct communication, and kindness.

• Translate clinical requirements into product strategy, requirements, and execution plans.
• Ensure clinical integrity, safety, and evidence alignment in product design and decision-making.
• Serve as connective tissue between Product, Care Delivery, and Clinical Leadership.

Headspace provides access to lifelong mental health support. They combine evidence-based content, clinical care, and innovative technology to help millions of members around the world. Headspace values include Make the Mission Matter, Iterate to Great, Own the Outcome, and Connect with Courage.

• You will define, build, and evolve foundational systems that enable autonomous agents to operate reliably in production.
• You’ll explore new approaches, prototype quickly, and turn what works into durable platform foundations.
• You’ll identify high-leverage architectural improvements, abstractions, and guardrails that expand what the platform can do while keeping it reliable, secure, observable, and maintainable under real-world conditions.

Kindo is an agent automation platform for DevOps and SecOps teams, helping organizations automate high-friction operational work using autonomous agents. They are a small, highly technical team with strong customer traction and real enterprise revenue, where engineers have direct ownership over critical systems.

• Own GitLab's AI control plane end-to-end, giving organizations visibility, control, and confidence as they adopt AI across the DevSecOps lifecycle.
• Focus on establishing a clear product strategy for AI Guardrails and Governance, translating customer demand into a prioritized roadmap, and delivering foundational capabilities.
• Define and drive the governance model for AI across GitLab, including hierarchical policy controls, feature-level toggles, role-based access, data-handling settings, and model-selection preferences that give organizations the granularity they need.

GitLab is an open-core software company developing the most comprehensive AI-powered DevSecOps Platform, used by more than 100,000 organizations and enabling everyone to contribute to and co-create the software that powers our world. Their high-performance culture is driven by their values and continuous knowledge exchange, enabling their team members to reach their full potential while collaborating with industry leaders to solve complex problems.

• Design and implement infrastructure to support LLM-based autonomous agents capable of multi-step reasoning, planning, and task execution.
• Architect and maintain cloud-native platforms that support end-to-end AI workflows, from model experimentation to high-availability production deployment.
• Implement security controls against prompt injection and ensure PII/PHI de-identification within agentic data flows.

Lirio is a technology/software company that provides expertise in a variety of behavioral science domains, data science, and machine learning to drive consumer engagement, close gaps in preventive and chronic care, and promote health and well-being. They are using a behavior change AI platform to deliver Precision Nudging health interventions.

• Design, train, and evaluate state-of-the-art VLA models for robotic systems.
• Implement scalable architectures for multimodal model fusion, continual learning, and domain adaptation
• Collaborate across engineering, robotics, and mission teams to integrate ML pipelines with onboard autonomy

Scout AI develops Fury, the first robotic foundation model for defense, to give U.S. forces overwhelming, adaptable, and autonomous power across every domain. They are a startup with a mission that demands urgency, precision, and relentless work, offering a rare opportunity to architect the future of defense.