Capital, but no desk
The money and the strategy are ready. Nobody on the team has built a trading desk.
QuantOps Architect
I build the systems quant desks run on, from research to execution.
For prop firms, family offices and quant teams: research first, then execution, risk and reporting as the desk grows. Two decades of production trading systems, including the core framework that ran $40M+ of capital across multiple accounts.
2004
In the markets since
7
Production frameworks
$40M+
Capital run on the core framework
1
Architect, start to finish
Problems I solve
Where quant teams get stuck
Backtests don't match live
Strategies that look good in research behave differently once real money is on.
Research lives on one laptop
Notebooks and scripts that only one person can run, and nobody else can check.
Pieces that don't fit
Data vendor, backtester and broker API bought separately, held together by hand.
What gets built
Three phases, each ending in something running
The first build is the smallest desk that can trade for real, with up to three months from agreement to first live trade. It grows from there, without being rebuilt.
Phase 01
Research
Data ingestion, storage and a backtesting environment. You can test a strategy yourself, on your own data, before any money moves. Available on its own, including an audit and overhaul for teams that already trade.
Phase 02
Execution
Order management and broker connectivity. An order reaches the market, comes back filled, and the position is where it should be.
Phase 03
Risk and reporting
Position, P&L, limits and a daily report. The desk trades inside its limits and leaves a record of what it did.
How I work
One architect, no handoffs
-
Ownership
One person designs and builds the pipeline. There is no integration seam between teams, because there are no teams.
-
Cadence
A working demo every two weeks, not a status report.
-
Your repository
Source, tests and documentation go into your repository.
-
Your server
A shared machine for the team on your infrastructure: JupyterLab, VPN access, one environment for everyone.
-
Handover
Runbooks, deployment notes and a walkthrough, so the desk does not depend on me being reachable.
Proof
What I've built
Three systems built on my code. All mid-frequency, all on real money.
2018 – 2022
Multi-account algorithmic fund
Problem: A fund trading futures, options and US stocks across several accounts had no single system from research to execution.
Build: A backtesting engine for ML models, the datafeed, multi-account execution via Trading Technologies FIX and Interactive Brokers, a Jupyter research environment over VPN, and the Linux servers.
Outcome: The framework ran $40M+ of capital across multiple accounts.
2023 – 2026
Swarm Investments LLC
Problem: A prop group needed signal research and execution on US futures.
Build: My execution framework adapted for them, a tool that generates swarms of alpha signals, FIX execution to Trading Technologies, and their Linux server operations.
Outcome: Live trading on US futures through TT FIX.
2021 – now
My own book
Problem: My own capital needed the same research and execution standard I build for clients.
Build: An execution framework (multi-leg positions, internal netting, smart orders) and a research engine (universal datafeed, alpha library, Monte Carlo portfolio risk).
Outcome: It runs my own and family money, and every client build starts from it.
Open source
Scope
What I don't do
-
Staff augmentation
No day-rate contracting and no borrowed developers.
-
Advisory only
No architecture documents that stop at the document.
-
Alpha and strategy research
You own the thesis. I build the process that tests it and the machine that runs it.
-
Performance claims
No published returns and no track records. That is not what is being sold.
Risk
What happens if I disappear
The desk keeps trading without me. Source, tests and documentation go into your repository, and every account is in your name. Runbooks cover every operational task, so your team can run the desk.
Contact
Tell me what you're building
Tell me which markets and frequency, who is on the team, and what exists today. If it's not a fit, I'll say so.