Razvan Zegheanu
Founder at Undisclosed Startup
Romania
Summary
👤
Senior🎓
Top SchoolRazvan Zegheanu is a seasoned software engineer and founder with 16 years of experience building scalable, distributed systems and six years focused on hosting highly distributed ML workloads. He led and optimized internal machine translation hosting at Amazon, helping scale services by orders of magnitude and cutting translation costs by more than half. Comfortable spanning backend systems, ML model hosting, and optimization, he has contributed to influential open-source ML projects such as Theano and pylearn2, fixing core library bugs and implementing optimization algorithms. Razvan is practiced at bridging cross-continental teams and design opinions to deliver pragmatic solutions for complex problems. Based in Romania, he combines deep hands-on engineering with an appetite for exploring new technologies and applying research-grade ML techniques in production.
16 years of coding experience
13 years of employment as a software developer
Bachelor's degree, Artificial Intelligence, Bachelor's degree, Artificial Intelligence at Polytechnic University of Bucharest
English
Github Skills (19)
algorithm10
python10
optimizers10
machine-learning10
deep-learning10
optimisation10
logistic-regression10
linear-algebra10
numerical-computing10
theano10
optimization10
code-optimization9
algorithms9
numerical-optimization9
numeric8
Programming languages (1)
Github contributions (5)
lisa-lab/DeepLearningTutorials
Jan 2010 - Mar 2013
Deep Learning Tutorial notes and code. See the wiki for more info.
Role in this project:
ML Engineer
ML EngineerContributions:57 commits in 3 years 3 months
Contributions summary:Razvan primarily contributed to the development of deep learning models within the repository. They implemented logistic regression with conjugate gradient descent and stochastic gradient descent optimization methods, demonstrating proficiency in Theano. Furthermore, they developed a multi-layer perceptron (MLP) model, including the use of L1 and L2 regularization. The user's work also involved modifying the dataset loading procedure, adapting the code for various batch sizes, and adding validation and test error reporting to provide evaluation metrics for the models.
deep-learning
lisa-lab/pylearn2
Jan 2010 - Dec 2012
Warning: This project does not have any current developer. See bellow.
Role in this project:ML Engineer
ML EngineerContributions:686 commits in 2 years 11 months
Contributions summary:Razvan primarily focused on implementing and improving machine learning algorithms within the PyLearn2 framework. Their contributions include the implementation of the linear conjugate gradient optimization method and its associated test, along with the integration of the minres algorithm. The changes also show modifications to existing cost functions and model definitions, indicating a focus on model training and optimization.
javascripttypescript
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.