Hark

Build the evaluation, data, and reinforcement-learning infrastructure that helps researchers improve multimodal and language models at scale. This hands-on role owns the systems that measure model progress and make training repeatable and reliable.

Hark is hiring an Application Security Engineer to own the security of its backend services, mobile apps, integrations, and emerging LLM-powered agent systems. This hands-on role combines secure code review, vulnerability remediation, threat modeling, and security tooling.

Build Hark’s native AI desktop experience for macOS and Windows, bringing multimodal intelligence, voice, vision, memory, and agent actions directly into users’ workflows. You’ll own core desktop features end to end and collaborate closely with research, model engineering, design, and hardware teams.

Build real-time embedded firmware and software for haptic and motion-control systems, shaping how next-generation AI hardware feels and responds. This hands-on role combines motor control, bench characterization, rapid prototyping, and close collaboration with product and industrial design teams.

Build hardware-in-the-loop test benches and software infrastructure that verify device software on real hardware, including radios, sensors, audio, cameras, and connectivity systems. This role owns automated test coverage, failure analysis, and integration with continuous testing workflows.

Build the digital worlds and scalable infrastructure that enable AI agents to learn, navigate software, and interact with dynamic environments. This full-stack role spans simulations, agent-gym infrastructure, distributed execution, and RL training and evaluation.

Hark is hiring a Graphics Engineer to own the embedded display and graphics stack powering its next-generation AI hardware. The role focuses on rendering engines, UI frameworks, GPU drivers, and high-performance graphics on resource-constrained platforms.

Own the deployment of advanced audio algorithms onto embedded DSP hardware, transforming desktop prototypes into optimized, real-time production features. This hands-on role spans algorithm redesign, low-level performance tuning, acoustic validation, and shipping audio functionality.