Valerio Carpani
Robotics Engineering Senior Manager at FlexLink, a Coesia company
Bologna, Emilia-Romagna, Italy
Summary
🤩
Rockstar🎓
Top SchoolValerio Carpani is a robotics engineering leader with 8 years of hands-on experience building production computer vision and deep learning systems for automated picking and intralogistics. Currently Senior Manager at FlexLink after leading its Robotic Lab, he previously led vision engineering at Fizyr where he moved models from research into robust, deployable pipelines. His open-source contributions include improving mAP evaluation for the well-known keras-retinanet project, reflecting attention to evaluation accuracy and practical tooling. Trained as an electronic engineer with study at Bologna and an exchange at Twente, he blends academic rigor with pragmatic product delivery. Colleagues value his ability to translate complex vision algorithms into maintainable systems and to spot subtle evaluation bugs that materially improve model reliability.
8 years of coding experience
6 years of employment as a software developer
Master's degree, Exchange student, Master's degree, Exchange student at University of TwenteLaurea Magistrale LM, Ingegneria elettronica, Laurea Magistrale LM, Ingegneria elettronica at Alma Mater Studiorum – Università di Bologna
Italian, English
Github Skills (7)
object-detection10
keras10
computer-vision10
machine-learning10
tensorflow10
python10
callback9
Programming languages (4)
Github contributions (5)
fizyr/keras-retinanet
Jun 2018 - Feb 2020
Keras implementation of RetinaNet object detection.
Role in this project:
ML Engineer
ML EngineerContributions:41 commits, 14 PRs, 37 pushes in 1 year 8 months
Contributions summary:Valerio primarily contributed to improving the mean Average Precision (mAP) calculation for the RetinaNet object detection model. They fixed bugs in the evaluation scripts, refactored code related to evaluation callbacks, and incorporated weighted average calculations. Additionally, the user addressed typos, corrected code style, and made the weighted average computation optional, indicating a focus on enhancing the accuracy and usability of the object detection framework.
deep-learningretinanetcomputer-visionobject-detectiontensorflow
fizyr/dr_ros
Jul 2019 - Jan 2021
Contributions:1 review, 1 PR, 3 pushes in 1 year 6 months
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.