Aurenium Capital

The Problem

Risk teams were reconciling price feeds, regulatory filings, and insider-transaction data by hand across fragmented sources. By the time a risk score was compiled, it was already stale against the market that had moved underneath it.

Off-the-shelf platforms modelled market risk in isolation from regulatory and insider-activity signals. What was needed was one pipeline that fused all three into a single automated, nightly-refreshed score, with reporting that ran itself.

Constraints

• The nightly retraining and rescoring cycle had to finish before markets opened, across two time zones.

• No single point of failure in the data pipeline — a missed AlphaVantage pull couldn't silently freeze a fund's risk score.

• Daily reports had to be genuinely personalised per institution, not a templated PDF with a name swapped in.

• Subscription entitlement had to update instantly on lapse or upgrade, not on a manual billing review.

• Every automated score needed an audit trail a compliance team could actually interrogate.

How Nexolve Built It

Aurenium runs as an automation chain rather than a single model: n8n orchestrates ingestion, an LSTM scores risk overnight, and reporting fires itself off the moment scoring completes, no analyst in the loop for the routine cycle.

1. 1

   Unified data pipeline

   n8n workflows pull equity, ETF, and commodity prices from AlphaVantage alongside regulatory filings (insider transactions, IBC events) every day, transforming and landing everything in Supabase Postgres without an analyst touching a spreadsheet.

2. 2

   Nightly LSTM risk scoring

   A model trained on 10 years of Indian and GCC market data re-scores every tracked position each night on a multi-factor risk index, so the desk opens each morning to numbers that reflect the last close, not last week.

3. 3

   Automated daily reporting

   Personalised risk reports, position alerts, and weekly opportunity digests go out via Nodemailer without a template being manually filled — each report is built from that institution's live portfolio state.

4. 4

   Entitlement-aware billing

   Stripe webhook handlers update access in Supabase in real time, so a lapsed subscription loses dashboard access instantly rather than at the next manual audit.

5. 5

   Explainable attribution

   Beyond a single score, the platform surfaces which positions, trades, and fees are driving P&L divergence — the automation has to justify itself to a compliance team, not just output a number.

Tech Stack

Outcomes

• Risk scoring runs nightly with no analyst compiling it by hand, 99.9% model accuracy on backtested scoring.

• Daily risk reports and position alerts generate and send themselves, personalised per institution.

• Subscription access updates in real time off Stripe webhooks, no lag between payment and entitlement.

• Attribution is explainable down to position and trade level, built for a compliance review, not just a dashboard.

• Live at aurenium.pages.dev, open for institutional demo bookings.