Chief Architect, principal developer, Stream Analyze
Sep, 2016 - Present
At Stream Analyze, I have played a pivotal role in the development and evolution of SA Engine, a cutting-edge distributed main memory database for highly efficient computational queries, focusing on real-time data processing, high performance, and system scalability. I started as the first employee, doing everything, and grew the organization to 9 engineers with a very low turnover.
Designed array-processing capabilities for numerical models and designed rewrite and optimization rules (patent pending) making SA Engine beat Tensorflow Lite and OpenVino by 275% on average on single core CPUs.
Developed WebAssembly (WASM) support, allowing SA Engine to run in WASI and browser environments.
Spearheaded the development of SA Federation Services, supporting tens of thousands of edges producing more than 1 million events per second.
Designed and maintained end-to-end secure deployment pipelines for customer projects. Making sure we always tested SA Engine 20+ platforms on every commit to every branch with more than 20k workflow runs.
Built a tight integrations for Python and Java, giving me a good understanding on how python and java works under the hood.
Worked closely with clients to adapt and optimize our solution for real-world production environments.
Sole developer of SA Studio and its VSCode plugin, a web-based tool for real-time query visualization and distributed SA Engine management.
Implemented the building blocks that SA Federation Services would build upon. Making SA Engine utilize standard web technologies by implementing http client and server, including websockets and TLS security using ARM mbedTLS inside SA Engine.
Implemented our first deep learning runtime SA.NN.