Julius Hietala
Working on applied ML systems at the intersection of perception, control, and real-time inference. Focused on building ML that runs in production and integrates with software and control stacks.
Currently interested in Vision-Language-Action (VLA) models with real-time controllers in the loop, targeting practical hardware rather than idealized execution.
Projects
Building ML systems that run in production, with a focus on robotics, computer vision, and real-time inference.
Dynamic Cloth Folding Robot
A robot agent for folding pieces of fabric. The paper was a best paper finalist at IROS 2022.
Stray Robots
Stray Robots is a set of tools to help build 3D computer vision datasets and models
SAM 2 Visualizer
An interactive visualization tool for Facebook's Segment-Anything 2 (SAM2) model. Experiment with SAM2's segmentation capabilities through an intuitive UI.
React Native Llama
A React Native mobile app for running LLaMA language models using ExecuTorch. Run LLaMA models directly on device with a React Native UI.
YOLOv5 iOS Object Detector
I built an iOS object detection app with YOLOv5 and Core ML. Check out the tutorial to know how you can build one too!
YOLOv8 Image Classification App
Using Core ML from React Native
crabnet
A simple neural network implemented using Rust 🦀
More on GitHub
Check out more of my open source projects and contributions on GitHub
Skills & Expertise
Technologies and domains I work with to build production ML systems
Core Languages
ML Frameworks
Computer Vision
Systems & Tools
Research Areas
Contact
Interested in collaboration, research discussions, or have questions about my work?