Vruddhi Tolia

Architected and developed a multithreaded C++ risk-validation engine, leveraging lock-free ring buffers and pre-allocated memory pools, to screen equities and options orders in real time, achieving 157 μs end-to-end latency and preventing over $10M in potential losses from limit breaches.

Engineered high-performance, large-data processing platforms in C++ and Java, normalizing and sequencing millions of exchange messages per second with gap-detection algorithms to ensure reliable, low-latency delivery within a distributed computing environment.

Optimized exchange-feed ingestion by implementing busy-spin polling and cache-line alignment, pinning latency-sensitive threads to isolated CPU cores, which cut P99 tick-to-trade latency by 35% across the firm's equities trading platform.

Collaborated daily with traders, quantitative researchers, and infrastructure engineers within a small, highly technical team to translate complex trading requirements into rapidly deployed, production-ready software solutions.

Developed AI-powered developer tooling using LangChain agents, RAG, NLP, and LLM function-calling to automate code review and test generation, successfully automating 46% of the SDLC workload and addressing complex engineering challenges.

Designed and implemented continuous performance benchmarking architectures using flame graphs, perf profiling, and custom microbenchmarks, proactively identifying hot-path regressions and fostering a culture of engineering excellence.

Led end-to-end development of TestPlan, a Python/Bash operational-stability platform, automating regression and integration testing, which increased test coverage by 40% and significantly reduced production incidents for critical trading infrastructure.