Suyash Sharma

I'm building high-performance C++ infrastructure, from an in-memory time-series database reaching 16M inserts per second to an ultra-low-latency order book simulator with strict price-time priority matching.

During my cybersecurity software engineering internship, I developed a C++17 anomaly detection engine for real-time log analysis processing over 100K logs per second. I also engineered crash-safe sharded persistence, an LRU cache, zero-copy parsing, and lock-free queues, reducing latency by 40%.

I work across distributed systems, operating system internals, memory optimization, and hardware-level profiling. My projects include a tensor and autograd engine, plus research on using Order Flow Imbalance to predict short-term market movements.