Ivan Fioravanti
Founder at Devocracy srl
Milan, Lombardy, Italy
Summary
🤩
Rockstar🎓
Top SchoolIvan Fioravanti is a founder and technology leader with 13+ years building scalable cloud and AI-enabled products, currently advising and chairing technology at CoreView while founding a local stealth AI startup in Milan. He combines deep hands-on expertise in Microsoft Azure, MongoDB, Kubernetes and OpenAI integrations with a track record of creating NoSQL-backed, HTML5-first enterprise solutions from his 4ward to CoreView journey and a successful 2018 exit. Ivan is an active investor and board member, helping grow the Italian startup ecosystem while translating strategic vision into engineering execution. His open-source ML engineering work includes practical enhancements to MLX examples—improving model management, LoRA dropout, CPU efficiency and colorized token output—showing a willingness to optimize both developer UX and model performance. A Mensa member and avid reader, he brings intellectual curiosity and a habit of continuous learning to complex technical and business challenges. He’s known for blending product-minded leadership with low-level technical contributions that accelerate adoption and operational efficiency.
13 years of coding experience
26 years of employment as a software developer
Liceo Scientifico "Luigi Cremona"
Italian, English
Stackoverflow
858reputation
42kreached
15answers
12questions
Github Skills (16)
machine-learning10
python10
command-line-interface9
pytorch8
githubaction-workflow7
lora7
github-ci7
huggingface7
windows6
azure6
php6
mongodb6
amazon-web-services6
xaml6
windows-runtime6
Programming languages (13)
Github contributions (5)
ml-explore/mlx-examples
Dec 2023 - Mar 2025
Examples in the MLX framework
Role in this project:
ML Engineer
ML EngineerContributions:8 reviews, 12 PRs, 83 comments in 1 year 3 months
Contributions summary:Ivan primarily contributed to the MLX framework example code. They implemented new features, such as colorized output for generated text based on token probabilities. They added an argument for max sequence length and model management functionality including scanning, retrieving information, and deleting models. Furthermore, the user introduced a dropout parameter to the LoRA configuration, improved the model's efficiency with CPU utilization, and corrected code indentations and naming conventions.
mlx
ivanfioravanti/azure-quickstart-templates
May 2015 - Oct 2023
Contributions:17 pushes in 8 years 6 months
azure
Find and Hire Top DevelopersWe’ve analyzed the programming source code of over 60 million software developers on GitHub and scored them by 50,000 skills. Sign-up on Prog,AI to search for software developers.