Florin Blanaru
Senior Software Engineer at Axelera AI
Bucharest, Romania
Summary
🤩
Rockstar🎓
Top SchoolFlorin Blanaru is a Senior Software Engineer with 11 years of experience specializing in compilers, ML/AI, custom ML accelerators and managed runtimes such as the JVM/OpenJDK. He builds end-to-end compiler stacks and heterogeneous execution layers—recently working on TVM-based compilers at Axelera AI and contributing to Apache TVM during his time at OctoML. At The University of Manchester he advanced TornadoVM with a PTX GPU backend, overlapping compute and data transfers, and redesigned memory management for high-throughput workloads. Florin combines low-level code generation (OpenCL/PTX) and bytecode manipulation with performance tuning for low-latency, parallel GPU/CPU systems, and he has hands-on experience integrating these optimizations into production CI and test infrastructure. Based in Bucharest, he brings a research-informed engineering approach that bridges academic projects and product-focused ML accelerator tooling.
11 years of coding experience
5 years of employment as a software developer
Master of Philosophy - MPhil, Computer Science, Master of Philosophy - MPhil, Computer Science at The University of Manchester
Github Skills (18)
tvm10
pytest10
python10
annotations10
testing10
bytecode-manipulation10
java10
annotation-processing10
javas10
parallel-computing10
machine-learning9
opencl9
deeplearning-ai9
compiler-compiler9
deep-learning9
Programming languages (5)
Github contributions (5)
beehive-lab/TornadoVM
Oct 2019 - May 2022
TornadoVM: A practical and efficient heterogeneous programming framework for managed languages
Role in this project:
Back-end Developer
Back-end DeveloperContributions:37 reviews, 325 commits, 32 PRs in 2 years 7 months
Contributions summary:Florin primarily contributed to the TornadoVM project by implementing the ASM library within its own module, focusing on bytecode inspection and the handling of parallel annotations. Their work included removing deprecated JVMCI methods and adding filtering for `@Parallel` annotations. The user also worked on the low-level code, particularly the OpenCL code generation and its connection to memory.
practicalgpgpumulti-coreprogramming-languagefpga
apache/tvm
May 2022 - Jan 2023
Open deep learning compiler stack for cpu, gpu and specialized accelerators
Role in this project:QA Engineer / Test Automation Engineer
QA Engineer / Test Automation EngineerContributions:39 reviews, 27 commits, 36 PRs in 8 months
Contributions summary:Florin's commits primarily focus on adding and modifying tests within the TVM repository, with a strong emphasis on Python unit tests. Their contributions include adding a utility to replace direct calls to pytest.main, indicating a focus on test infrastructure and organization. The changes span various test files, including those related to meta-schedule, relay, and other core TVM functionalities. The user is also responsible for running the test cases using the CI pipeline.
metalvulkancompilertensoropencl
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.