Making code provably faster - Gen AI, Performance Optimization, Code Generation, Formal Verification, Symbolic Execution, Static Analysis, Metaprogramming, Functional Programming, Applied Category Theory

AI/ML Systems, AI Agents, Large Language Models, workflow orchestration, data platform

Designed and implemented back-end distributed infrastructure for financial flow and account identification computations. Working on High Performance Computing, distributed algorithms and systems, system optimization, scalability, graph computations, computational linear algebra, blockchain, using Spark, Pandas/Numpy, NetworkIt, GraphFrames, AWS.

Staff software engineer and architect for Grammarly's global data platform. Design and implementation of machine learning and analytics Feature Store (including support for low latency feature engineering pipelines). Worked on data platform query DSL and back-end. Used Spark, Scala, Cassandra, Redis, Zeppelin, AWS, Databricks.

Machine learning engineer for geolocation intelligence. Created deep learning and classic models for travel, commuting and home identitication with vehicle geolocation data, enriched with demographic data, using TensorFlow, scikit-learn, Python, Vertica, SQL. Also worked on outlier detection, deep learning for time series, and statistical models.

Senior data platform and machine learning software engineer for Apple Software Engineering and the Core Operating System (kernel, Unix) group. Designed and implemented a complete ETL process and data analytics platform for automated build and test results for Core OS and all of Apple Software Engineering, based on Splunk, Kafka and Cassandra. Launched and implemented a major data analysis and visualization project for test data of all Apple Software Engineering. Analyzed test data using machine learning and statistical models while leading various projects: failed test log clustering and analysis; bad test detection; faulty device detection; test stability and reliability analysis; test timeout optimization. Assisted engineering managers and engineers throughout Apple efficiently query, analyze and visualize their data.

Lead engineer and architect for the test automation and tools team, creating the back-end of the Core Operating System (kernel, Unix) automated root (from code change) building, mastering and testing system, which was then adapted by all of Apple Software Engineering. Some of the subsequent features designed and implemented: a new (12 times faster) automated embedded root mastering and packaging solution; a re-queuing algorithm to balance lab device queues and prevent device starvation; a system-wide mechanism to override defaults for restoring and testing devices; an automated chrooted build system to build and test code changes; the first fully automated Core OS Nightly Build testing system.

White box test code architect and developer for the Virtual Memory (VM) subsystem and File Systems. Designed and implemented a complete general low-level kernel system call unit test framework, with numerous Mach VM unit tests. Ported ztest (user space emulation of the kernel for testing the ZFS file system) from Solaris to Mac OS X. Wrote automated file system performance benchmarking suite for Mac OS X and Windows. Performed complete FileVault corruption vulnerability analysis.

Mac OS X 10.4 (Tiger) French localization team lead. Set up the first Apple internal continuous localization process. Localization engineering lead for iPhone, iTunes, iPod, Intel CPU, Boot Camp. Started translation automation and similarity engine projects, using LSA and clustering.