Test & Improve AI Systems

Editor’s noteRed teaming seat — you create prompts to break AI models and document where they fail against safety taxonomies; light on dev, heavy on judgment work.

Editor’s notePrompt and eval work shaping e-commerce AI training data — you build realistic shopping simulations and score model outputs, not ship production code.

Editor’s noteCreative technologists design and maintain generative AI video workflows in a film production setting — heavy on workflow building, moderate Python, and heavy AI fluency.

Editor’s noteQuality-specialist seat evaluating AI agent behavior at scale — you're watching for hallucinations, looping, and escalation failures, then feeding that back to improve how agents perform.

Editor’s noteCustomer-facing AI adoption role — you translate AI capabilities into working systems for Fortune 500 clients, with hands-on work in prompt design and workflow configuration, not backend engineering.

Editor’s noteAdversarial AI testing role — you probe AI systems for safety and security weaknesses through prompt injection, attack chains, and adversarial inputs.

Editor’s noteJunior AI quality and validation role — you test AI outputs, document findings, and flag issues before release; minimal coding, heavy on structured QA work.

Editor’s noteAudio annotation and linguistic QA role for Grok's voice capabilities — you judge speech quality and cultural nuance, not build AI systems or workflows.

Editor’s noteRed team role that stress-tests frontier AI models for dangerous outputs — you bring offensive security craft, not engineering infrastructure, to a structured evaluation program.

Editor’s noteRed teaming seat evaluating AI model safety — you construct prompts and document where models fail, not write code or ship features.

Editor’s noteQA lead role defining and executing quality standards for AI prompt outputs — you own the evaluation lifecycle, not the underlying models or infrastructure.

Editor’s notePM role owning agent fleet product and chat flow evals — prototype fast with AI tools, run 15-20 user sessions weekly, decide what ships.

Editor’s notePart-time contractor seat contributing senior ops expertise to AI training data — you design real-world tasks and evaluate how AI handles operations scenarios.

Editor’s noteOwn the systems and metrics that determine whether AI models actually work in the real world—golden sets, eval pipelines, accuracy SLAs, and automated QA.

Editor’s noteEnablement role scaling AI adoption across an edtech company — you build the playbooks, prompts, and training, not the platform. Light on dev, heavy on making teams actually use AI day-to-day.

Editor’s notePM seat owning agent behavior end-to-end — you define how AI agents reason, interpret intent, and execute workflows, then measure whether they actually work.

Editor’s noteStrategic AI enablement seat — you own the roadmap, governance, and champion networks that make the rest of Headspace AI-capable, not the AI products themselves.

Editor’s noteSupervise live AI agents in a call center context — step in when agents stumble, catch failure patterns, and feed what you learn back to the team. Light on dev, heavy on judgment and real-time oversight.

Editor’s noteBuilder-first PM role prototyping and shipping AI adoption products — vibe coding with Claude Code/Cursor is non-negotiable, but you're shipping features, not production engineering.

Editor’s noteTrust & Safety role building and running AI workflows that flag content, route reports, and handle escalations — you're measured on automation rate and handle time, not on shipping features.

Editor’s noteOperations role configuring and improving Mercury’s customer-facing chatbot and email agents — you monitor performance, expand bot scope, and optimize workflows rather than build the underlying platform.

Editor’s noteContract eval/annotation role for AI output quality — you judge whether AI-generated content meets accuracy, cultural, and brand standards, with flexible hours and no coding required.

Editor’s noteEntry-level AI ops seat watching dashboards and logs to catch AI failures before they escalate — you flag and coordinate, engineers do the fixing.

Editor’s noteRevenue systems lead role that builds and runs AI-powered automations with Clay, Zapier, and n8n — you're owning the GTM tech stack and its workflows, not just advising on it.

Editor’s noteConversation design role writing prompts, optimizing chatbot flows, and QA-testing LLM experiences for customer journeys.

Editor’s noteEvaluation role owning RAG and agent quality frameworks, with research-grade Python instead of production infrastructure.

Editor’s noteHealthcare QA role reviewing AI agent calls, catching errors, labeling issues, and improving real workflows.

Editor’s noteOperator seat embedded in revenue, platform, or research work, using frontier AI to tighten execution rhythms.

Editor’s noteQA role reviewing AI support conversations for accuracy, policy fit, tone, and recurring failure patterns.

Editor’s noteRed-team role finding jailbreaks, ranking model failures, and documenting attack paths so safety teams can patch them.

Editor’s noteExpert prompt role owning AI character behavior, personality systems, and evaluation frameworks for social generative AI experiences.

Editor’s notePrompt-focused AI role owning use-case translation, agent behavior oversight, and quality testing for commercial real estate workflows.

Editor’s noteFound on Sierra career page.

Editor’s noteAxion is looking for a PM specifically for 'Agents', requiring prior experience in agentic or automation-heavy products.

Editor’s noteJerry.ai's 'Agentic AI' PM role is a clear fit for the mission, focusing on LLM systems within a personal assistant context.

Editor’s noteHigh-signal role for evaluating the next generation of AI video models.

Editor’s noteEnterprise-focused AI role applying agents and RAG to business process automation.

Editor’s noteRecently posted role for Adobe seeking an AI Eval Specialist.

Editor’s noteHamilton AI is a pure vibe-coding role. Applied Labs is a hands-on AI agent product role.

Editor’s noteStrong fit for Jerry.ai - focus on agent pm.

Editor’s noteStrong fit for EvenUp - focus on ai enablement/adoption.

Editor’s noteStrong fit for Airtable - focus on workflow designer.

Editor’s noteAn interesting operational role for someone interested in the evaluation and quality side of AI agent deployments.

Editor’s noteSelected for focus on Conversation Designer and AI fluency.

Editor’s noteSelected for focus on Ops and AI fluency.