Technology & platform
Engineering for Markets That Never Sleep.
We treat the trading platform like a product: low-latency market access, a petabyte-scale research lake, GPU training clusters and a model-serving layer built for microseconds.
Where engineers thrive
We Build Almost Everything In-House.
Our team functions more like a software company than a traditional financial firm. The goal is simple to state and hard to do: the most capable platform for learning systems in the industry.
Take Ownership
A flat organisation and small teams mean engineers own systems end to end, from design to on-call, and see their work in production within weeks.
Build What Doesn't Exist
Point-in-time feature stores, microsecond model serving and exchange connectivity are not off-the-shelf problems. We build them, and we measure them.
Ship to Production
Engineers sit with the model teams they support. Feedback arrives in minutes, not in quarterly reviews, and success is measured in live performance.
Full stack. Global impact.
One Platform. Research to Execution.
The same platform that trains a model overnight serves it in production the next morning. Data, compute, risk and execution are designed together, so nothing is lost in translation between research and the market.
- Proprietary market access, data, research, compute and compliance technology
- Colocation in the major market centres we trade
- One codebase from backtest to live, with identical risk checks on both
Key details
The Platform in Numbers.
Differentiated frameworks
Our Stack.
Innovation extends well beyond trading and risk technology. These are the layers we invest in most.
Data
Versioned, immutable datasets on Arrow and Parquet, streaming ingestion at millions of events per second and a feature store with point-in-time correctness.
Compute
Kubernetes-orchestrated research clusters, GPU pools for training and inference, and compute on demand for every researcher.
Languages
Python for research and tooling. Rust and C++ on the fast path, where we are investing strategically in Rust for safety-critical systems code.
Model Ops
A model registry, canary deployments, continuous evaluation against live data and automated drift detection tied to our kill switches.
A team of technologists
Diverse Skills. One Commitment to Performance.
Our engineers come from exchanges, hyperscalers, hardware companies and research labs. What unites them is a taste for hard systems problems and a habit of measuring before believing.
Qualities we prize
What Our Best Engineers Have in Common.
Drive
Making an impact and working until the job is actually done, not merely handed off.
Communication
With fellow engineers and with the researchers and traders our systems serve.
Collaboration
Within small, agile teams and across the firm, where the best idea wins regardless of its origin.
Passion
For innovative approaches to complex systems, and for the craft of making them fast, correct and observable.
Culture
A Culture of Excellence.
A results-oriented environment with supportive colleagues, regular tech talks, hackathons and the chance to see your code move markets.
Reach your greatest potential
Help Us Build the Platform.
We bring the platform, the capital and the research culture. You bring the ideas. Take the next step to learn about our teams, our hiring process and life at Balboa Line.