About the job
About Hark
Hark is at the forefront of artificial intelligence, developing innovative, personalized intelligence systems that are proactive, multimodal, and capable of engaging with the world via speech, text, vision, and persistent memory.
We are merging this intelligence with cutting-edge hardware to forge a universal interface that connects humans and machines. Unlike conventional AI that operates through outdated chat boxes and devices, Hark is pioneering the future with agentic systems that engage naturally with individuals and their environments.
Our mission is to create multimodal models and next-generation AI hardware, meticulously designed as an integrated interface for a new era of intelligent systems.
About the Role
As an Integration Engineer, you will take ownership of the integrations layer that empowers Hark's AI agent to be genuinely effective in real-world applications. Your responsibility will be to connect our AI to the tools, services, and data sources that users rely on daily.
Your work will involve designing and implementing the protocols, connectors, and execution primitives that enable Hark to perform real actions, such as managing emails, calendars, and third-party API interactions. The true potential of our agentic system hinges on its ability to connect and act.
This role offers significant ownership within a compact team where you will collaborate directly with platform engineers and model researchers. The integrations you develop will set the parameters for Hark's capabilities.
Responsibilities
- Design and construct the integration layer connecting Hark's agent to third-party services, including email, calendars, productivity tools, communication platforms, and developer tools.
- Oversee the complete lifecycle of integrations: from authentication flows to API clients, schema normalization, action execution, and robust error handling. Ensure reliability beyond just the happy path.
- Establish and implement tool and action protocols that allow the agent to engage with external systems in a secure and consistent manner. Facilitate the seamless addition of new integrations.
- Collaborate closely with platform engineers to ensure that connectors are integrated smoothly into the agent runtime, ensuring fast, observable, and appropriately sandboxed tool execution.
- Work in partnership with model researchers to define tool schemas and action semantics that models can reliably interpret, as the quality of the interface directly influences model behavior.
- Instrument integrations for reliability and accuracy, tracking failures, surfacing regressions, and ensuring rapid feedback loops with the team.
- Identify high-leverage integration opportunities to maximize impact.

