Senior Software Engineer II

Invoca is hiring a Senior Software Engineer II for its Integrations team. The role builds and operates customer-facing integrations and AI-powered features, owns projects from design through production, and improves the reliability and development practices of the team.

Responsibilities:

• Own a project, or a major piece of one, from scoping through shipping, and confirm afterward that it achieved the outcome it was meant to produce
• Recognize where your work touches other systems and teams, and resolve those dependencies and conflicts rather than handing them off
• Define what good looks like for the work you own and hold that bar, rather than waiting for it to be defined for you
• Own technical design for a feature or service. Decide as clearly what to leave out as what to build, and back your choices with a working prototype
• Weigh testability, scalability, and maintainability without being prompted, and favor the simplest approach that will hold up
• Use AI agents as a routine part of writing production code: run several in parallel, get productive quickly in unfamiliar parts of the codebase, and review AI output efficiently by focusing on the risky parts
• Build your own skills, agents, or tooling when it would make the team faster, and improve the parts of our shared development system your work depends on
• Know where it is safe to let AI run and where a human must review before anything reaches production, and set that bar for the work you own
• Build customer-facing features on top of AI systems: handle non-deterministic outputs, design and run evaluations, manage latency, and debug unexpected model behavior
• Establish safe, observable, and auditable ways to bring AI into the product, with monitoring and tracing in place so the team can see how agents behave with real users
• Understand the customer's real problem before building starts, and confirm afterward that it was solved
• Take part directly in customer feedback sessions, and use product and adoption metrics to guide what you build next
• Partner with your Product Manager to prioritize based on measurable customer and business impact, not technical elegance alone
• Troubleshoot independently across the team's services, contribute to post-incident reviews, and recommend process and monitoring fixes alongside code fixes
• Spend review time where the risk is highest, and identify gaps in review and verification tooling
• Lead team discussions, ask for differing opinions before a decision, and support the team's direction once it is set
• Communicate clearly with cross-functional partners, and help teammates get better at working with AI through pairing, demos, and documentation

Requirements:

• 5+ years of professional software engineering experience, with a track record of measurable impact in production environments
• At least one full year of actively using AI tools and agents to write production code, as a core part of your daily workflow rather than an occasional assist
• Experience shipping customer-facing, AI-powered product features with real users: working with non-deterministic outputs, evaluations, latency, and debugging unexpected model behavior. Building models yourself is not required
• Experience building your own skills, agents, or developer tooling that made a team faster
• Sound judgment about when to let AI run and when a human must review before anything reaches production
• A product-engineer mindset: you can take an open-ended problem, find the right signal from customer data and product partners, and ship a solution that solves the customer's real issue
• Comfort across the stack, with depth wherever the team leans
• An eagerness to automate. You look for chances to offload work to AI rather than needing to be convinced
• Strong technical judgment with a bias toward simplicity and adoption over theoretical completeness
• The ability to connect technical work to customer and business outcomes, and to communicate that connection clearly
• RAG, evaluations, agentic products, chatbots and conversational bots; evaluating and improving AI agents with monitoring, tracing, and simulation data
• building MCP tools, agentic pipelines, background jobs and async processing (Kafka or similar), third-party integrations
• Ruby on Rails (our stack), TypeScript/JavaScript (where we are moving), MySQL
• any other primary language
• You have shipped a third-party integration and operated it afterward: OAuth, webhooks, vendor APIs, retries, idempotency, and a way to know it failed before the customer reported it
• You have built or run an MCP server or another tool-calling integration that an agent used with real customers, including keeping one customer's credentials and data separate from another's, and a way to check that the tool call was the right one