Software engineer
I am a software engineer with 15 years of experience across cloud infrastructure, online gaming, and financial services. I like practical systems work: event-driven services, reliable data flows, performance tuning, and enough machine learning to keep things interesting. I am currently seeking EU roles with work visa sponsorship.
Leading the Tidal team on data ingestion and synchronization services. Built foundational AWS services with Lambda, DynamoDB, Kinesis, and SQS, and helped shape event-driven architecture, clean architecture, and CQRS practices.
Worked on Oracle Cloud Infrastructure compute services, including placement, distributed auditing, anti-entropy, integration testing, performance monitoring, GPU passthrough, and low-level hardware control systems. Promoted from Intermediate Engineer to Senior Engineer, then Principal Member of Technical Staff.
Built and optimized gaming server and client systems in C++, Node.js, and TIScript. Optimized a game simulator to run roughly ten times faster and built ZeroMQ-based server messaging.
I prefer software that has a clear job: move data correctly, use resources carefully, recover from failure, and stay understandable when more people join the project.
A small Python language-model experiment, useful as a compact way to revisit the mechanics behind sequence prediction.
A generative adversarial network implementation from first principles, written to understand the moving parts rather than hide them.
A C++ speed comparison around fast inverse square root, reflecting a long-running interest in practical performance questions.
A JavaScript library supporting AI development in games, tied to interests in algorithms, simulation, and decision-making.
Partial MSc thesis material on neuroevolution and machine learning.
Why a strange hypothesis deserves a careful reading before it deserves dismissal.
Some AI safety work is just ordinary engineering caution applied to unfamiliar systems.
Why small personal tools can be valuable even when they are not portfolio-grade products.
Speculation is useful, but it becomes stronger when it exposes itself to observation.
Notes on implementing simple ML systems to build intuition before depending on frameworks.
A short argument for keeping performance work empirical, boring, and close to the workload.