Levente Hunyadi
Data Architect Principal Engineer
Budapest, Hungary
Summary
👤
Senior🎓
Top SchoolLevente Hunyadi is a Data Architect and Principal Engineer with over a decade of experience designing and shipping data-intensive, distributed systems across industries from ed-tech to healthcare and media. He combines rigorous academic training (PhD-level work in parameter estimation and model reconstruction) with hands-on leadership, having built central data warehouses, real-time analytics platforms, and high-availability ML-driven services. Levente excels at translating ambiguous business needs into production-grade architectures while mentoring cross-functional teams of engineers and product leaders. His open-source ML contributions include adding GRU support and improving compatibility in the well-regarded frugally-deep library, reflecting a practical focus on model portability and robustness. Comfortable across multiple paradigms and languages, he navigates the trade-offs between short-term delivery and long-term platform health. Based in Budapest, he brings a track record of lowering operating costs and operationalizing research into scalable products.
10 years of coding experience
12 years of employment as a software developer
Doctor of Philosophy (PhD) parameter estimation and model reconstruction problems in the context of errors-in-variables systems, Doctor of Philosophy (PhD) parameter estimation and model reconstruction problems in the context of errors-in-variables systems at Budapest University of Technology and Economics
English, German, Hungarian
Github Skills (13)
neural-network10
keras10
machine-learning10
c-language10
recurrent-neural-networks10
convolutional-neural-networks10
deep-learning10
tensorflow10
cprogramming-language10
python10
cpp9
unit-testing9
edge-computing7
Programming languages (4)
Github contributions (5)
Dobiasd/frugally-deep
Nov 2018 - Jul 2019
A lightweight header-only library for using Keras (TensorFlow) models in C++.
Role in this project:
ML Engineer
ML EngineerContributions:13 commits, 9 PRs, 19 comments in 7 months
Contributions summary:Levente primarily contributed to the implementation and improvement of machine learning model functionalities within the `frugally-deep` library. They added support for the GRU (Gated Recurrent Unit) layer and fixed errors in various compiler implementations, ensuring compatibility with Keras models. Significant effort was also dedicated to adding comprehensive unit tests and integrating causal padding for Conv1D layers, demonstrating a focus on model accuracy and robustness. The user also worked on transforming kernel matrices from CuDNN format to Keras format for broader compatibility and usability of the library.
pythonkeras-modelsc-plus-pluskeras-tensorflowheader-only-library
hunyadi/pysqlsync
Sep 2023 - Feb 2025
Contributions:4 PRs, 220 pushes, 17 branches in 1 year 5 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.