Development · Quant & trading software

Trading software, built by people who trade.

Bots, backtests, integrations and desk tools for funds and trading businesses. We test them on the real venue, then hand over the code.

Who it's for

You have the strategy. You need the engineers.

Generalist dev shops build apps, not trading logic, and freelancers rarely stand behind the money handling. We sit in between: trading know-how, an Asian time zone, and code that has already run on real venues.

Small funds and prop teamsIn Singapore, Hong Kong, Australia, the Gulf or Europe, with an edge to automate and no engineering team.
Desks outgrowing their toolsPast what no-code bot platforms can do: cross-venue, options, hedging or custom risk.
Physical-asset and FX businessesGold, pawn and remittance operators who need pricing, hedging and book-keeping tools of their own.
What we build

Built for real money, because ours runs on it.

Each of these is something we first built for our own desks, not a slide-deck capability.

Strategy bots

Basis and carry, options selling, market making and cross-venue spreads, with dry-run modes and kill switches from day one.

Execution

Order routing, chase and TWAP, with correct handling of partial fills, minimum sizes and contract units.

Exchange & broker integrations

REST and websocket connections to crypto exchanges and FX brokers, with money and units named in every field.

Risk & hedging engines

Delta hedging, inventory hedges and exposure limits, written and tested to run unattended.

Pricing tools

Quote engines like a gold melt-value pricer and an FX hedge calculator, fed by live market data.

Dashboards & internal tools

Cross-venue scanners, trading dashboards, trade books and back-office screens your team will actually use.

Research that can say no

We report a dead idea as dead.

Backtests charge real fees, funding and slippage, then the idea gets a paper run before any capital. Plenty of ideas don't survive that, including several of our own. You get the result either way, written up plainly.

AI agents in the workflow

An assistant that knows your book.

We wire LLM agents into trading and operations through MCP servers, so your team can ask questions and record work in plain language. The rules that move money stay in code.

Read the bookPositions, hedge status, inventory and alerts, answered from live data instead of a spreadsheet.
Quote and recordPrice a deal, then log it, with a preview of what the write will do before anything is saved.
Guardrails in codeLimits and validations live in the system, not in the prompt, so the agent can't talk its way past them.
How an engagement runs

Fixed scope, proven live, then handed over.

Small team, a direct line to the people writing the code. If you'd rather we run it afterwards, see Infrastructure.

01

Scope

Venues, instruments, risk rules and what "done" means, agreed in writing before building.

02

Build & test

Tested code, with money and contract units named everywhere, because that is where trading systems break.

03

Prove it live

Paper trading and dry-runs against the real venue, then small size, before full capital.

04

Hand over

Code, tests, docs and deploy scripts are yours, with an optional support retainer.

Stack

Boring tools, used carefully.

We pick proven technology and spend the effort on correctness, not novelty. We build your systems; we don't take or manage outside money through this service.

Languages
Python for research, execution and services; TypeScript for dashboards and web apps.
Data
Postgres and SQLite for books and ledgers; streaming market data over websockets.
Venues
Major crypto exchanges for spot, perpetuals and options, plus FX brokers.
Delivery
Docker images, test suites that run before every deploy, and written docs.

Have a system you need built?

Tell us what you're trading, where, and what's in the way. We'll tell you plainly whether we can help and what it would take.

Get in touch
Manila, Philippines · desk running