RJ Ascani
Software Engineer at Meta
Seattle, Washington, United States
Summary
🤩
Rockstar🎓
Top SchoolRJ Ascani is a software engineer based in Seattle with eight years focused on embedded systems and TinyML, now contributing to Meta after a multi-year tenure at Google working on TensorFlow Lite Micro. He brings deep firmware experience from roles at Impinj and Motorola, developing RAIN RFID readers and mission-critical radio systems. At Google he improved low-power ML deployment by adding int8/float SVDF support and refining TFLite memory and type handling, demonstrating an ability to reconcile ML algorithms with tight resource constraints. His open-source contributions touch flagship projects like TensorFlow, where he fixed memory management, type mismatches, and refactored code to reduce compiler warnings—work that improves reliability across broad embedded deployments. Practical, detail-oriented, and comfortable at the intersection of systems software and machine learning, he reliably turns complex, resource-constrained requirements into production-ready solutions.
8 years of coding experience
18 years of employment as a software developer
Bachelor of Science (B.S.) Information Technology and Computer Science, Bachelor of Science (B.S.) Information Technology and Computer Science at University of Miami
English
Stackoverflow
1reputation
0reached
0answers
0questions
Github Skills (20)
c-language10
memory-management10
machine-learning10
refactor10
tensorflow10
sys10
embedded10
refactoring10
cprogramming-language10
continuous-deployment9
ml-deployment9
python9
deeplearning-ai9
deep-learning9
optimization9
Programming languages (3)
Github contributions (5)
tensorflow/tflite-micro
Apr 2022 - Apr 2022
Infrastructure to enable deployment of ML models to low-power resource-constrained embedded targets (including microcontrollers and digital signal processors).
Role in this project:
ML Engineer
ML EngineerContributions:394 reviews, 1 commit, 118 PRs in 1 day
Contributions summary:RJ's contributions focused on enhancing the tflite-micro framework for low-power embedded targets. Their work involved adding support for int8 and float SVDF implementations for non-HiFi Xtensa cores, enabling more flexibility in model deployment. Furthermore, the user removed dead code and optimized the framework by removing unused code sections, like the `Dequantize` op and `LSTM` vector usages. Also, they have improved testing by initializing variables in the test output and fixing the benchmarking tools.
signalml-modelslow-powerprocessorsdeployment
tensorflow/tensorflow
Oct 2023 - Jul 2024
An Open Source Machine Learning Framework for Everyone
Role in this project:Back-end Developer
Back-end DeveloperContributions:8 reviews, 7 PRs, 14 comments in 8 months
Contributions summary:RJ's contributions primarily involved modifying and improving the TensorFlow Lite (TFLite) library. They addressed memory management issues by conditionally disabling specific functions based on a static memory definition. The user also corrected type mismatches and misalignments within the codebase. Additionally, the user refactored code to prevent compiler warnings and align integer array types with flatbuffer data types.
pythondata-sciencedeep-learningmlmachine-learning
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.