IT Infrastructure Engineer - AI
About the job
- Able to work Monday through Friday in the UK or Spain (Barcelona) time zones.
- At least 5 years of experience with cloud infrastructure and/or DevOps engineering, ideally with direct AWS and/or GCP experience in compute, networking/VPCs, IAM, and storage.
- Practical experience deploying and hosting applications across AWS or GCP cloud infrastructure.
- Hands-on experience with AI concepts such as MCP, AI agents, and other AI-related technologies.
- Strong scripting and automation skills in Python and/or Bash, working knowledge of Git for version control, and comfort with JSON/YAML for configuration and API work.
- Experience with containerization using Docker or Podman and container orchestration such as Kubernetes, ECS, or Cloud Run, including running and troubleshooting containerized services.
- Experience with infrastructure-as-code tools such as Terraform, CloudFormation, or Pulumi and CI/CD pipelines such as GitHub Actions.
- Familiarity with API-based integrations, including managing keys, scoping permissions, reading documentation, and troubleshooting connectivity issues.
- Ability to train and mentor IT staff on infrastructure and DevOps practices relevant to the role.
- Familiarity with identity provider tools such as Okta and with OAuth and SSO concepts from an infrastructure and security perspective.
- Ability to learn in a constantly shifting environment where change is unavoidable.
- Set up, configure, and maintain MCP (Model Context Protocol) servers that connect AI systems to internal tools and data sources, including scoping access, managing credentials, and troubleshooting integrations.
- Design, provision, and support hosting environments for internally built AI applications, from lightweight PaaS deployments to full cloud infrastructure on AWS or GCP, including VPCs, compute, containers, and load balancing, based on application complexity and scale requirements.
- Provision and support AI agent and application infrastructure, including service accounts, API key management, permission scoping, rate-limit monitoring, and access auditing for agents operating across internal systems and cloud environments.
- Build and maintain CI/CD pipelines and infrastructure-as-code to support repeatable, secure deployment of AI-hosted applications and supporting services.
- Maintain documentation for AI integrations and hosted environments, including architecture diagrams, access maps, runbooks, cost and usage tracking, and troubleshooting guides.
- Stay current on AI tooling and the cloud ecosystem, and identify tools, platforms, or integrations that could benefit the IT team or wider organization.
- Support administration of the company’s IT cloud infrastructure (AWS and/or GCP) and DevOps toolset as they relate to IT’s AI infrastructure.
- Become a subject matter expert for AI-hosted platforms, cloud integrations, and supporting infrastructure.
- Monitor uptime, performance, and cost of hosted AI applications, and implement alerting and observability to detect issues before they affect users.
- Identify, prioritize, and address technical debt across cloud infrastructure, deployment pipelines, and integrations.
- Partner with IT system administrators and Engineering and Security teams to improve reliability, automation, documentation, and supportability of hosted environments.
- Act as an escalation point for complex infrastructure and hosting issues for the IT support team.
- Collaborate with cross-functional teams, including Security, Engineering, and People Operations, to improve cloud governance, security posture, and adoption of AI tool and application hosting.
- Act at times as a project lead, managing stakeholders and collaborating with non-technical counterparts; demonstrate solutions and work with business owners on AI build and support work.
- Document procedures, system changes, and troubleshooting workflows.
- Familiarity with AI features in existing SaaS platforms such as Atlassian Rovo, Slack AI, Okta AI, ChatGPT, and Google Gemini.
- Experience with MCP servers or similar tooling.
- Cloud or AI platform certifications, such as AWS, Google Cloud, or Anthropic certifications; hands-on experience is valued more highly.
About the company
UserTesting provides a platform for gathering customer feedback to guide the development and delivery of digital products and experiences. It uses an on-demand sourcing engine and a global, opt-in audience network to deliver rapid, high-quality usability and experience feedback throughout the product development cycle. Unlike many research tools, it emphasizes its extensive experience-research capability, enabling early input in the design process and reducing guesswork with direct human insights. The service is subscription-based, serving a wide range of clients from small businesses to large enterprises. The company’s goal is to help teams build better digital experiences by continuously collecting and applying user feedback across development stages.
Company Size
1,001-5,000
Company Stage
IPO
Headquarters
San Francisco, California
Founded
2007
Get referred to UserTesting
See people who can refer or advise you
Simplify's Take
What believers are saying
- January 2026 acquisition of User Interviews added 6 million participants and premium B2B recruiting.
- September 29, 2026 Auros rebrand signals demand beyond research, into AI training and evaluation.
- Eric Johnson said customer demand outruns supply, supporting near-term revenue expansion through 2026.
What critics are saying
- July 2026 layoffs hit UserTesting, and Revelio shows headcount fell 23.7% since 2023.
- Product overlap between UserTesting, User Interviews, and Auros invites channel conflict and integration drag.
- OpenAI, Google, and specialist eval vendors commoditize AI feedback, crushing pricing power.
What makes UserTesting unique
- Auros pairs UserTesting insights with User Interviews recruitment across 34 countries.
- ARQ scores AI agents across five dimensions, linking research to model evaluation.
- UserTesting for Figma and MCP servers embed validation directly inside builders' workflows.
Help us improve and share your feedback! Did you find this helpful?
Benefits
Remote Work Options
Growth & Insights and Company News
6 month growth
0%1 year growth
0%2 year growth
0%UserTesting has rebranded as Auros, positioning itself as "the human network behind AI" as it expands beyond customer-experience research into AI training and evaluation. The company now serves large enterprises and labs building frontier models. CEO Eric Johnson reports the company's vetted network has grown from approximately 1.5 million to over 7.5 million people in two years. Auros has launched ARQ, a new tool that scores AI agents across five dimensions. Johnson indicates customer demand currently exceeds supply.
UserTesting names Darren Ong as Regional Sales Director. UserTesting has named Darren Ong as Regional Sales Director, adding a new sales leadership role as the company continues to build its presence across regional markets. Ong shared his move to UserTesting, marking a new step in his sales career with the customer experience and human insights technology company. In his new role, he will focus on regional sales activities and customer relationships as UserTesting continues to serve organizations across its global markets. UserTesting provides a platform that enables Product, Marketing, Digital, and Customer Experience teams to gather feedback and human insights to support decisions around products, services, and customer experiences. The company combines participant feedback, video-based research, and AI-powered analysis as part of its platform. The appointment comes as customer experience technology continues to place greater emphasis on combining quantitative data with direct feedback from customers. For sales organizations operating in this market, regional leadership can play a role in developing customer relationships, understanding local business requirements, and connecting organizations with experience research capabilities. UserTesting has maintained a regional sales structure across markets including North America, EMEA, and Asia Pacific. In 2024, the company appointed dedicated general managers for North America and EMEA and a regional vice president for Asia Pacific as part of its global sales organization. Ong's appointment adds to this regional focus and places him within UserTesting's broader sales organization. His role will be relevant to the company's engagement with businesses looking to incorporate customer feedback and human insights into product, marketing, digital, and CX decisions. The move also reflects the broader shift toward customer insight platforms that connect research with business and marketing workflows, giving sales teams another area in which to engage organizations evaluating customer experience and research technologies.
UserTesting announced the launch of new MCP Servers. Search. Two new Model Context Protocol (MCP) servers enable teams to recruit participants, create studies, and validate AI-generated ideas from within ChatGPT, Claude, Figma Make, and other AI applications - spanning a combined network of more than seven million participants across 34 countries UserTesting, the leading provider of customer insights for the enterprise, today announced the availability of its Model Context Protocol (MCP) Servers for both UserTesting and User Interviews platforms, giving product, design, UX, and research teams direct access from within AI applications including ChatGPT, Claude, Figma Make, and other MCP-compatible clients. The new capability embeds customer insight directly into AI-powered workflows, allowing teams to recruit participants, create studies, launch tests, and retrieve customer feedback without leaving the tools where ideas are generated. Instead of switching between disconnected AI assistants, research platforms, and collaboration tools, teams can move seamlessly from concept to customer validation in a single workflow. "AI has dramatically accelerated how teams generate ideas, designs, and content. Speed only creates value when it's grounded in real customer understanding," said Jennifer Artabane, Vice President of Product Management at UserTesting. "With MCP servers for both UserTesting and User Interviews, we're bringing human insight directly into the AI workflows where decisions are made. Teams can move from asking AI what's possible to validating what actually works with real customers without breaking their creative flow." As organizations increasingly rely on AI to accelerate product development and design, one challenge remains: AI trained on existing patterns tends to pull toward the familiar, not the distinctive. Building products that stand out requires insight into customer emotion, tastes, and discernments - and that insight comes from real people, not inference. The UserTesting and User Interviews MCP servers close that gap by connecting AI-generated concepts with fast, real human feedback. This draws on UserTesting's enterprise-grade testing platform and User Interviews' participant network. "At Picsart we've built an AI-powered product development lifecycle that gets from concept to shipped experience in a fraction of the time it used to take," said Oli Mival, VP Product Strategy, Innovation and Experience. "UserTesting's MCP server lets us keep real people in that loop without slowing it down, so human insight is woven through every stage rather than a checkpoint we pass once. It means we're not just building fast; we're building the right things in the right way." Using the UserTesting and User Interviews MCP servers, teams can: * Recruit participants directly from within supported AI applications * Create and launch studies without leaving their AI workflow * Bring customer insights back into ChatGPT, Claude, Figma Make, Gemini, and other MCP-compatible clients * Validate AI-generated concepts with real people before launch * Turn raw sessions into synthesized findings, with every takeaway anchored to the evidence behind it * Move from idea generation to customer feedback in a continuous workflow, regardless of which platform's participant network is best suited to be the audience Human Insight Where Decisions Happen The introduction of these MCP servers is a response to a broader shift in the product development lifecycle in the age of AI. The product development lifecycle is compressing: research and customer feedback are no longer a discrete phase teams move through once, but an ongoing input woven throughout a much shorter cycle. That shift changes what is required of customer insight itself. It has to be accessible at every stage of development, not gated to a research phase. It has to be just as easy for AI agents to reach as it is for humans, since agents are now doing a growing share of the work that once required a person to open a separate tool. And it has to sustain the same quality and rigor regardless of who, or what, is asking the question. By embedding UserTesting and User Interviews into AI-assisted workflows, teams spend less time switching between applications and more time iterating on ideas informed by real customer feedback. Whether prioritizing what to build next, weighing a business trade-off, or refining a user experience, teams can validate decisions with customers before moving forward. The launch also complements UserTesting's broader strategy of embedding customer understanding across the product development lifecycle, enabling organizations to move faster while maintaining confidence that AI-assisted decisions reflect real customer needs, not assumptions. Both MCP servers are currently available in Early Access and support ChatGPT, Claude, Figma Make, Gemini, and other applications compatible with the Model Context Protocol. Additional capabilities for surfacing customer insight throughout AI workflows will continue to expand over time.
UserTesting has launched Model Context Protocol servers for its UserTesting and User Interviews platforms, enabling teams to access customer insights directly within AI applications like ChatGPT, Claude, and Figma Make. The servers allow users to recruit participants, create studies, and validate AI-generated concepts without leaving their AI workflow, drawing on a combined network of over seven million participants across 34 countries. The integration addresses a challenge in AI-accelerated product development: ensuring speed is grounded in real customer understanding rather than pattern-based inference. Teams can move from concept generation to customer validation in a continuous workflow. "AI has dramatically accelerated how teams generate ideas, designs, and content," said Jennifer Artabane, Vice President of Product Management at UserTesting. "With MCP servers for both UserTesting and User Interviews, we're bringing human insight directly into the AI workflows where decisions are made." The MCP servers are currently available in Early Access.
UserTesting has launched Advanced Targeting, a feature that integrates the User Interviews participant network directly into its platform. This marks the first major product integration since UserTesting acquired User Interviews. The new capability allows research, product, design, and marketing teams to recruit verified B2B professionals and niche consumer audiences from a single workflow. Users can define advanced audience criteria, set custom incentives, and manage recruitment through a shared prepaid balance. Advanced Targeting provides access to 3.2 million professionals across 140 industries. The feature supports unmoderated tests, interaction tests, live conversations, and surveys. It is available as an early access offering through 30 September 2026. The integration represents a step towards unifying UserTesting's and User Interviews' participant ecosystems, which together comprise over seven million participants across 34 countries.