Senior Software Developer, AI Agents

Knak is a mission-driven company Why? Because our time is limited, our competition is fierce, and our margin for error is small. To have the biggest impact on the world, we stay focused on our core mission, which is... Empowering marketers to be creative . That’s why Knak exists. In a world where AI has made everyone a creator, Knak helps marketers turn ideas into on-brand, high-performing campaigns at scale. The world's most AI-forward companies, including Meta, OpenAI, NVIDIA, Vercel, Notion and Stripe, use Knak as their agentic marketing production layer. Here's what they and our investor Insight Partners figured out: creation got easy, but that’s making production harder. Every new idea and brief still has to be built, versioned, checked, approved, and shipped across every channel, and that work runs on handoffs, review cycles, and a lot of copy-paste. That middle stretch of the marketing stack has no system, no standard, and no owner. Knak is the flexible factory for that work. Our customers set the brand rules, choose the tools that feed in and the platforms that send, and decide where AI does the work and where people do. We handle the repeatable parts, so marketers get their time back for the parts that aren't. We're built by marketers, for marketers. We started with email because it's the hardest channel to get right and where many campaigns win or lose. Now we're bringing that same production discipline to every channel our customers use. The Role: Senior Software Developer, AI Agents We’re looking for an experienced software developer to build and scale the AI agents behind Knak’s platform. In this role you'll develop the agent harness itself: the orchestration, tool calling, context management, and guardrails that let agents analyze data, generate complex marketing assets, and deliver real value to our customers. You'll also ensure we know they work, by building the evaluation suites that measure agent quality and catch regressions before they reach customers. This is a hands-on role for someone who wants to be deep in agent internals every day. You'll also contribute to architecture decisions, mentor teammates, and collaborate across engineering and product to shape the future of marketing automation. What You’ll Be Doing Build the agent harness : Design and implement the core agent runtime in Python: orchestration loops, tool definitions and execution, memory and context management, structured outputs, and failure handling. Own agent evaluations : Build and maintain eval datasets, automated evaluators, and LLM-as-judge pipelines (LangSmith) so every prompt, model, or harness change is measured, not guessed at. Iterate on agent quality : Diagnose agent failures from traces, tune prompts and tool designs, and run model comparisons and benchmarks to drive measurable improvements. Create AI microservices : Build Python and Node.js services for agent hosting and LLM interaction. Build reliable…