Three pains every sales and BD team hits weekly. Each one is what your reps actually complain about, not what a feature page would call them.
Pre-trade prep eats two hours a morning across funding rates, open interest, FRED calendar, and last week's journal, and the setup is already half gone when you finally size.
Position sizing drifts to gut feel after three winning days, then one 4 percent loss on a single name eats the month because Kelly never got a second look.
Post-trade review never happens on a busy week, so the same overfit setup gets re-entered next Monday and you cannot say what changed.
Every pipeline below is a shape you wire on the canvas using the crew and tools further down. Not a feature we ship for you, a pattern you configure.
QuantAnalyst pulls funding, OI, and the FRED 30 day calendar, MacroEconomist context is held in Static context so the regime call stays consistent, ReportWriter compresses to one page. Brief lands in the trader's inbox before the open.
QuantAnalyst kicks off a fresh melaya_backtest_start on the current six month window, then pulls the sweep results to check parameter sensitivity. Cross-run memory carries the last sweep so the trader sees what drifted week over week.
ChiefRiskOfficer runs parametric and historical VaR plus CVaR, replays Flash Crash, COVID, and FTX scenarios, and prints the exact circuit-breaker level. HITL gate blocks the order stage until the trader acknowledges the worst case in dollars.
QuantSkeptic applies Deflated Sharpe for multiple-testing correction, names one failure mode, and returns APPROVED, CONDITIONAL, or REJECTED. The rag_retrieve tool pulls the contradicting paper or the last losing trade on the same setup.
PortfolioManager shows the Kelly fraction with the formula and the dollar size, ExecutionSpecialist picks the order type and prints fill-rate and slippage in basis points. The order routes through melaya_exec_sim_place_order first, then waits behind pause_for_human for the live click.
After the trade closes, ReportWriter pulls the fill log, ExecutionSpecialist compares realized to expected slippage, QuantSkeptic flags whether the thesis or the execution failed. Replay reproduces every step so next Monday's setup starts honest.
Real personas from the quant_firm crew. Each ships with a tuned system prompt and a default tool allowlist. Swap models per persona on the canvas.
Scores the signal with Sharpe and its standard error, runs parameter sensitivity, and applies Bonferroni or Benjamini-Hochberg correction before any thesis ships.
Decomposes slippage into spread, market impact, and adverse selection in basis points, with fill-rate estimates per order type and latency band.
Sizes the trade with fractional Kelly at 0.25x to 0.5x, runs Ledoit-Wolf shrinkage on the correlation matrix, and shows the math on every recommendation.
Calculates parametric and historical VaR plus CVaR, runs Lehman, Flash Crash, COVID, and FTX stress scenarios, and sets exact circuit-breaker thresholds.
Applies the Deflated Sharpe Ratio for multiple-testing correction, blocks any thesis without walk-forward validation, and returns APPROVED, CONDITIONAL, or REJECTED with the exact conditions.
Compresses the five specialist sections into one decisive pre-trade brief with ACTION, CONFIDENCE, and recommended params on page one.
Every tool below is a real shared tool from the Melaya bundle. Allowlist per agent; HITL-gate the writes; revoke any of them in one click.
Pull funding, open interest, liquidations, OHLCV, and the order book for the setup the trader is about to take. Read only, no HITL needed because nothing writes.
Run a fresh backtest or pull a parameter sweep so QuantAnalyst can verify the thesis on this regime before sizing. Read and start jobs only, no live execution path from this bundle.
Stage the live order the trader will click and route the paper version through the sim account first. melaya_place_order, melaya_cancel_order, and melaya_set_leverage are HITL by default for this crew.
Cross check perp positioning across venues with funding, OI, liquidations, and the long-short ratio. ExecutionSpecialist uses this to size the adverse selection cost. Read only.
Pull equities, FX, and macro context for the discretionary trader running futures or single names alongside perps. Read only, no HITL needed.
Reach FRED for rates, DXY, inflation prints, and the 30 day release calendar that moves position sizing. Read only, every chart cites its series id.
Pull 10 K, 10 Q, 8 K, and Form 4 data straight from EDGAR for the equity trader who refuses to size before reading the filing. Read only.
Build the per workflow store from your trade journal, win and loss postmortems, playbook, and research papers. Powers all three knowledge layers. Indexing is HITL gated when new corpora are added.
General research, structured fetch, and a Slack drop for the post-trade review. pause_for_human is the explicit approval node any workflow can call before order_send.
Every pipeline ships with three layers of knowledge access. Mix and match per agent on the canvas. No shared vector space with another tenant, no surprise reads, no opaque retrieval.
includeContextPer-pipeline documents appended to specific agents' input on every run. The ICP brief, playbook, pricing sheet, or won-deal email corpus. Whatever needs to be there before the agent thinks. You pick which personas get which docs.
rag_retrieveA scoped tool granted per-agent. When the agent decides it needs more depth, it queries the workflow's vector store on demand. Same knowledge base as Static context, accessed only when the model asks for it.
pipeline_memoryPipeline-level state that carries from one run to the next. Yesterday's research is in scope for today's follow-up. The crew remembers what it already prospected, what got approved, what was sent. The audit log is the second-order knowledge base.
Sandbox tier is free with no card. Join the waitlist and we will email you the moment a slot opens.
← Back to every use case