Julius Hietala

About

I'm a senior applied ML and robotics systems engineer focused on production systems at the intersection of perception, simulation, control, and real-time inference.

I build ML that survives contact with real software and hardware constraints: latency, data quality, deployment targets, controller integration, evaluation, and reliability. My work spans production ML systems, computer vision pipelines, robot learning infrastructure, mobile/on-device inference, and simulation environments for embodied AI.

I've shipped production ML systems, co-founded technical companies, written practical deep learning and computer vision tutorials, and published peer-reviewed robotics work at IROS and in the International Journal of Robotics Research. My research and engineering work includes visual feedback control, robotic cloth manipulation, sim-to-real pipelines, and system-aware model integration.

I'm currently most interested in applied AI systems for the physical world: robotics, VLA-style model integration, simulation-based evaluation, perception/control loops, and automation problems where models need to run reliably outside idealized demos.

Core Areas

• Applied ML and robotics systems
• Computer vision and perception pipelines
• Simulation, evaluation, and sim-to-real infrastructure
• Real-time inference and controller integration
• Research-to-production engineering
• Python, PyTorch, C++, Rust, Core ML, ExecuTorch, ROS

Experience

Publications