About Me
Experienced in developing and optimizing machine learning models on devices. Proficient in Python, Swift, CoreML, PyTorch, and MPSGraph, with expertise in iOS and macOS application development. Employed and optimized CoreML models for the Apple Neural Engine, integrating them with iOS and macOS applications. Developed tools for profiling Core ML models. Led deployment of fine-tuned LLM models with Nvidia GPUs and designed RISC processors. PhD in Engineering with a focus on Explainable AI.
Contact Information
fabio.guzmanf@upb.edu.co
Featured Projects
CLIP Finder
An iOS app for semantic offline searches of gallery photos using natural language descriptions or the camera, employing Core ML models optimized for the Neural Engine and MPSGraph for pre- and post-processing on the GPU.
View ProjectCore ML Profiler
A visual profiling tool for macOS to analyze Core ML models, providing estimation times and validation messages for the Neural Engine.
View ProjectPhD Thesis - Explainable AI
Developed attribution methodologies to accurately identify concepts in convolutional neural networks (CNNs), generating reliable explanations to enhance confidence, transparency, and security in decisions made by CNNs.
View ProjectNatalius RISC Processor
Designed and simulated a RISC processor, suitable for tape-out using SKY 130nm technology, including a hand-coded assembler-level compiler tool.
View ProjectCore ML Blog
Unlocking Core ML Potential: Mastering Async Batch Predictions
Explore the async and batch prediction in Core ML and learn how to optimize your app's performance.
Read MoreConverting Models to Core ML
Learn how to convert PyTorch models to Core ML format and deploy them on Apple devices.
Read MoreProfiling Core ML Models
Explore techniques for profiling Core ML models, both with dedicated tools and on-device, to optimize performance across Apple devices.
Read MoreIntroduction to MPSGraph
Discover the power of Metal Performance Shaders Graph for accelerating machine learning and numerical computing on Apple Silicon GPUs.
Read MoreEducation
Ph.D. in Engineering, Computer Vision - Explainable AI
Universidad Pontificia Bolivariana
2017 – 2022
Master of Advanced Studies in Embedded System Design
Università della Svizzera italiana - Switzerland
2012 – 2013
M.Sc. in Engineering
Universidad Pontificia Bolivariana
2009 – 2011
B.Sc in Electronics Engineering
Universidad Pontificia Bolivariana
1999 – 2005
Skills & Expertise
Languages
- Python
- Swift
- SwiftUI
- C
- C++
- Verilog
Frameworks & Tools
- CoreML
- MPSGraph
- PyTorch
- Xcode
- VSCode
- Git
- Linux shell scripting