Quantitative trading firm building ML-driven multi-asset investment strategies
Trexquant operates a quantitative trading platform powered by Python, C++, PyTorch, and Hugging Face, processing market data across equities, ETFs, and futures. The stack reveals a hybrid infrastructure: GPU-accelerated ML (PyTorch, Hugging Face, OpenAI) for signal generation paired with low-latency execution tools (FIX, Protocol Buffers, Kubernetes on AWS/GCP). Active hiring across research (25 roles) and engineering (30 roles) reflects concurrent scaling of systematic strategy development and platform modernization—execution platform revamp and low-latency trading system development top the roadmap.
Notable leadership hires: Quantitative Research Head
Trexquant is a quantitative investment firm founded in 2012, headquartered in Stamford, Connecticut. The firm develops machine learning models and trading algorithms across equities, ETFs, and futures markets using advanced statistical methods and a vast data architecture. Operations span the United States, Canada, China, and India. Core competencies center on multi-asset portfolio construction, backtesting infrastructure, trade execution optimization, and systematic strategy development. The organization emphasizes both research depth (dedicated quantitative research team) and engineering rigor (infrastructure and platform development).
PyTorch, Hugging Face, and OpenAI models for signal generation. The stack also includes Python, NumPy, and Pandas for data processing, alongside C++ and Kubernetes for low-latency execution and infrastructure scaling.
The firm posts roles in the United States, Canada, China, and India, reflecting geographic expansion of both research and engineering capacity.
Other companies in the same industry, closest in size