About MLC
Learn about our mission to democratize AI through machine learning compilation and systems co-design, enabling high-performance deployment of AI models.
Our Mission
MLC is dedicated to democratizing artificial intelligence by making high-performance model development and deployment accessible to everyone, everywhere, on any device.
We focus on compiler-driven and system-level innovations that unlock efficient AI across platforms. While we highlight modern LLM workloads, our work spans a broad range of models including vision, speech, and multimodal applications. Our goal is to empower developers, researchers, and organizations to deploy advanced AI—from laptops and phones to the cloud and edge—while maintaining exceptional performance, efficiency, and privacy.
Accessibility
Making AI accessible to developers and users across all platforms and devices.
Privacy
Enabling on-device inference to protect user privacy and data security.
Efficiency
Optimizing for minimal resource usage and maximum performance.
Our Journey
Project Inception
MLC-LLM project began as a research initiative to optimize large language model inference for mobile and edge devices.
First Release
Launched MLC-LLM v0.1.0 with basic inference capabilities and support for popular LLM architectures.
Mobile Support
Introduced MLC-LLM Mobile framework enabling native iOS and Android deployment with optimized performance.
Community Growth
Reached 1,000 GitHub stars and established active Discord community with 500+ members.
Production Serving
Released MLC-LLM Serve for production-ready deployment with auto-scaling and load balancing.
Major Optimization
Achieved 50% memory reduction and 2x inference speed improvement through advanced compilation techniques.
Enterprise Adoption
Major tech companies began adopting MLC-LLM for production workloads, reaching 10,000+ GitHub stars.
Current Milestone
MLC-LLM v0.3.0 with breakthrough performance improvements and expanded hardware support.
Core Team
Dr. Tianqi Chen
Core Team · Carnegie Mellon University
Creator of Apache TVM and XGBoost. Leads compiler and system innovations in MLC, enabling efficient deployment of modern AI workloads including LLMs.
Dr. Zhihao Jia
Core Team · Carnegie Mellon University
Researches systems for ML and high-performance inference/serving. Contributes to scalable execution and optimization across heterogeneous hardware.
Dr. Xupeng Miao
Core Team · Purdue University
Focuses on low-latency LLM serving and compiler-runtime co-design, including speculative decoding and token-tree verification techniques.
Join Our Team: We're always looking for talented developers, researchers, and contributors to join our mission. Explore our projects or check out our tutorials to learn more about our work.
Technology Stack
Core Technologies
ML Frameworks
Hardware Support
Our Partners
We're proud to collaborate with leading organizations and institutions that share our vision for democratizing AI.
Carnegie Mellon University
Research collaboration on machine learning compilation and optimization techniques.
University of Washington
Joint research on mobile AI and edge computing optimization strategies.
Purdue University
Collaboration on low-latency serving systems and compiler-runtime co-design.
Apache Software Foundation
Integration with Apache TVM ecosystem for advanced compilation capabilities.
Open Source Community
Collaborative development with the broader open source AI community.
Ready to Join Our Mission?
Whether you're a developer, researcher, or organization looking to leverage AI technology, MLC provides the tools and community support you need.