Bring trading analytics, execution review and scenario calculations into a programmable array runtime. l combines q/k expressions and qSQL with an execution model built around vector layout and memory traffic.
01 / The workload
Trades, positions, market observations and model outputs describe different views of the same business. Analysts need to compare those views across desks, instruments and scenarios, then follow a surprising aggregate back to the records that produced it.
Group normalized trade and position records by desk, book or instrument. Calculate notional measures and analyze supplied P&L components at the level needed for the investigation.
Work with sensitivities and outputs from your pricing and risk models. Use vector arithmetic and reductions to compare scenarios and locate the instruments driving a change.
Analyze prepared execution records and market benchmarks across client segments, venues and trading conditions. Study distributions as well as average outcomes.
Apply your analytical expressions to a defined historical sample. Compare calculated measures with existing references and inspect the conditions under which the results differ.
02 / From data to analysis
Keep supplied model results and the dimensions needed to explain them together through the analytical calculation.
Bring normalized positions, instrument attributes and scenario outputs with matching identifiers and valuation conventions.
Apply vector calculations to scenario columns, group by trading dimensions and reduce the results into an explainable view.
Return the summary and its contributing rows to the application used by analysts, with the assumptions needed to compare runs.
/ Compare scenario exposures l>risk:([]book:`rates`fx`rates;dv01:100 -50 200) l>select sum dv01 by book from risk book | dv01 -----|----- fx | -50 rates| 300 l>exec sum dv01*2 from risk 500j l>select from risk where dv01>100 book dv01 ---------- rates 200 l>
03 / Why l
You write q/k. The runtime considers the physical representation of the data, choosing encoding-aware operations where supported and materializing values when needed.
04 / Your first evaluation
Choose one desk or book, a reconciled position snapshot and a small set of scenario outputs with established reference totals.
Available as a research preview for local evaluation.
Why moving fewer bytes matters for analytical workloads.
Read the paper Integration guideCall l from Python, pass typed values and bring results back to your application.
Read the guide Integration guideConnect native services to l through the Rust IPC client.
Read the guideYour data. Your next question.