Customer Support Engineer

Pocket is a Y Combinator W26 company building a premium hardware plus software product in the AI voice recorder space. We are obsessed with quality and speed. We are scaling very fast in the US and now European markets and building a device experience that can become the category leader.

The engineer behind customer support. We run a self-hosted helpdesk with an AI agent on top of it, and you own that stack: the automations that take repetitive work off the team, the integrations that put order and device data in the ticket, the pipelines and dashboards that show where volume comes from. A growing share of the work is enterprise. We sell Pocket into companies now, and B2B support runs on different rails - named accounts, SLAs, admin tooling, security reviews - and most of that layer doesn't exist yet. This is a hands-on role. You sit with the support team every day and you ship your own code.

• Our self-hosted helpdesk: uptime, upgrades, configuration, and the patches we contribute back upstream
• Automations and integrations that remove manual steps - routing, macros, ticket enrichment from our order, fulfillment, and device systems
• The AI support agent and the knowledgebase behind it: retrieval quality, evals, and pushing resolution rate up
• The support data layer: pipelines, models, and dashboards covering volume, contact drivers, deflection, CSAT, and team performance
• Analysis that turns ticket data into decisions - what's breaking, what it costs, what product needs to hear
• Enterprise support infrastructure: named-account routing, SLA tracking, admin and fleet-level views, and the workflows behind security reviews and compliance requests
• Account-level reporting for enterprise customers - deployment health, fleet status, usage - built once and used by support, sales, and the customer
• Internal tooling for the CS team. If something is missing, you build it

• You've owned a service in production end to end
• Strong with TypeScript and SQL, and comfortable running and debugging a self-hosted app
• SQL well past the basics, and enough Python to do the analysis yourself
• You've built dashboards people use
• Hands-on with LLM tooling - prompts, retrieval, evals. You can tell why an AI agent gave a bad answer and fix it
• You've built for business customers and you know what changes at that tier: admin roles, audit trails, SLAs, single sign-on, data handling that survives a security review
• You work directly with non-technical people and turn vague operational problems into systems
• Bias to ownership. You find the manual step and delete it

Bonus points for:

• Time spent in or next to a support or ops team
• Enterprise support tooling: SLA engines, account hierarchies, SSO, SOC 2 or HIPAA workflows
• Shopify, helpdesk, or CRM API work
• Consumer hardware sold into companies

• Work directly with us and learn fast
• Direct impact on how the company operates day to day
• High-trust, high-responsibility environment
• Competitive compensation
• Scope to take on more as you earn it