Mihai Morariu
Software Engineer at Stellantis
Upper Merion Township, Pennsylvania, United States
Summary
🤩
Rockstar🎓
Top SchoolMihai Morariu is a software engineer with 11 years of experience, currently building products at Stellantis from Upper Merion Township, Pennsylvania. He combines hands-on backend and machine learning work, contributing to notable open-source projects like keras-retinanet and chainer-chemistry where he improved model implementations, fixed tricky data/adjacency bugs, and added tests. His background as a product engineer informs a pragmatic approach to shipping robust systems and improving developer workflows. Mihai’s contributions show an ability to bridge research code and production-ready software, particularly in object detection and chemistry-focused deep learning. Colleagues rely on him to refactor and harden codebases while keeping feature delivery on track. He brings a quietly persistent engineering style that favors correctness, reproducibility, and incremental improvement.
11 years of coding experience
7 years of employment as a software developer
Oakland University
Github Skills (16)
retinanet10
data-preprocessing10
model-building10
keras10
chemistry10
computer-vision10
machine-learning10
object-detection10
deep-learning10
tensorflow10
python10
chainer10
refactor9
mlp9
refactoring9
Programming languages (6)
Github contributions (5)
chainer/chainer-chemistry
Apr 2018 - Dec 2018
Chainer Chemistry: A Library for Deep Learning in Biology and Chemistry
Role in this project:
Back-end Developer & Data Scientist
Back-end Developer & Data ScientistContributions:92 commits, 26 PRs, 7 pushes in 8 months
Contributions summary:Mihai made contributions related to data preprocessing, dataset handling, and model training within the Chainer Chemistry library. They fixed issues related to dataset validation, particularly concerning invalid dataset types within the tox21 dataset. They also implemented changes to the underlying models by merging code from a related research branch and addressing specific bugs regarding how the adjacency matrix is handled within the RSGCN preprocessor. Furthermore, the user added unit tests for the MLP model and improved the provided training examples.
chemistrygraph-convolutional-networkspythondeep-learningbiology
fizyr/keras-retinanet
Oct 2017 - Oct 2017
Keras implementation of RetinaNet object detection.
Role in this project:ML Engineer
ML EngineerContributions:8 commits, 4 PRs, 7 pushes in 8 days
Contributions summary:Mihai contributed to the implementation and modification of the RetinaNet object detection model. Their work involved building submodels, fixing bugs in existing code, and merging changes from another branch. The user refactored code and renamed functions. Overall, the contributions focused on refining the model architecture and functionality.
deep-learningretinanetcomputer-visionobject-detectiontensorflow
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.