Faraz Intelligence

Overview

Designed and built a full-stack private equity portfolio analytics platform powered by Claude AI. Faraz replaces fragmented fund manager spreadsheets with a real-time pane of glass for covenant monitoring, automated LP reporting, and a natural language 'Ask Faraz' interface for querying fund data.

Objectives

• Replace fund manager spreadsheets with a single real-time source of truth
• Build 'Ask Faraz', a Claude-powered natural language interface for querying portfolio data
• Automate LP quarterly reporting and covenant breach early-warning
• Design a fund dashboard surfacing Net IRR, TVPI, DPI, NAV, and distribution waterfalls

Approach & Activities

• Designed the IA around PE mental models: fund → portfolio companies → debt schedules → LP commitments
• Built Ask Faraz using Claude API with streaming responses and structured fund data context
• Fund dashboard: IRR gauge vs benchmark, MOIC color-coded portfolio table, distribution waterfall, debt summary, and LP commitments on one screen
• Debt schedules view covers all tranches (Senior, Mezz, Revolver, PIK) with real-time covenant health indicators and maturity wall chart
• Stripe billing with trial, Starter, Growth, and Enterprise tiers, plus a demo provisioning flow
• Full Next.js App Router with Tailwind v4, Recharts, and Lucide icons

Challenges & Lessons

• Deeply nested PE data models (fund → vintage → company → tranche → covenant) required careful schema design
• Covenant breach detection needed to feel proactive, not reactive: early-warning logic surfaced before violations occurred
• LP Portal required presenting complex capital account data in a format non-technical LPs could understand

Impact