David Lariviere

Capital markets have long been the proving ground for new technologies, from low-latency networking and custom hardware to AI and advanced analytics. Finance firms move fast, iterate quickly, and have historically been first to adopt technologies that later reach the broader market.

As trading systems chase ever-lower latencies, ensuring fairness in market data and market access has become a complex technical challenge. While cable length equalization has dominated fairness debates, today’s concerns extend to switch architectures, cross-connects, port-level latency differences, and communication strategies—all of which can offer participants nano and picosecond-level advantages.

As enterprises expand their AI capabilities, MLOps platforms are being pushed to their limits—especially with the rise of large language models. LLMOps introduces new complexities, from managing prompt versioning and fine-tuning workflows to optimizing retrieval-augmented generation (RAG) pipelines and ensuring governance at scale.

Software development organizations have invested significantly in the Developer Experience (DX): tools, processes, and work environments to increase productivity, collaboration, and happiness of critical talent. But the Engineer Experience (EX) for FPGA teams typically lags that of software teams, even as the number of FPGA engineers on Wall Street continues to grow.

It's no secret that FPGAs are a favorite component for ultra-low latency system engineers. Firms can gain a significant edge by placing relatively simple trading strategies in an FPGA to avoid user space entirely.