Earlier this year at CppCon – the world’s premier gathering for advanced C++ practitioners – Tower Research Capital hosted an Ask Me Anything session. Engineers, quants, and curious attendees gathered to discuss how Tower approaches trading, infrastructure, and large-scale systems engineering.

The questions we received ranged from language choices and platform architecture to the realities of building ultra-low latency systems at global scale. The conversation made one thing clear: there’s strong interest in how firms like Tower blend cutting-edge engineering with systematic trading.

To extend that discussion beyond the conference floor, we’ve compiled some of the most frequently asked questions about our business and our answers here. Think of this as a distilled AMA – an inside look at how we build, operate, and evolve one of the world’s most sophisticated trading platforms.

How is Tower’s core engineering team structured, and how does it support trading?

Engineering is built around a hub-and-spoke model – our platform powers a diverse array of internal trading teams who work independently from one another. Core Engineering develops the shared infrastructure while our trading teams plug into that platform to research, build and implement their trading strategies. That balance lets us move fast on strategy and innovation while keeping critical systems consistent, stable, and continually optimized.

What languages do you use, and for what?

We use a blend of C++, Python, and Rust, each for a specific purpose:

• C++ powers our most latency-sensitive systems – market data, execution, and low-level infrastructure.
• Python is used heavily in research, data analysis, and orchestration.
• Rust is increasingly used for infrastructure and services where memory safety and robustness matter.

In all cases, we start with the end goal and work backward, selecting the language that best suits the task at hand.

What are Tower’s biggest technical challenges right now?

Scaling performance while keeping systems clean and easy to maintain is always the challenge. We’re constantly refining a global platform that supports very different trading styles – with ultra-low latency and data-intensive approaches – without fragmenting code. There’s also the organizational side: ensuring independent teams can innovate while still benefiting from shared libraries, infrastructure, and standards.

How do quants and engineers collaborate during the research-to-production lifecycle?

Quants and engineers collaborate across three primary lifecycle stages:

• Research and Prototyping – Quants test hypotheses, design features, and train models at small scale. This work runs on shared infrastructure (e.g., data layers, feature pipelines, experiment frameworks, and compute) built and maintained by research engineers. Collaboration in this phase focuses on enabling fast iteration and experimentation so quants can efficiently validate ideas.

• Productionizing – Quants and production engineers work closely to translate research code into production-ready systems. This includes optimizing feature computation, model inference, and overall system performance. Engineers primarily own deployment, runtime reliability, and market connectivity (data feeds, order routing, execution systems) required to run strategies live.

• Post-Production and Iteration – Quants and engineers jointly monitor live performance across both model and infrastructure dimensions. They diagnose issues such as latency regressions, inefficient feature computation, data discrepancies, execution slippage, etc. Insights from this phase feed back into model refinement by quants and ongoing infrastructure improvements by engineers.

Watch this space for more questions and answers. To learn more about career opportunities at Tower, explore here.