Kova

Overview

Designed and built an AI-native algorithmic trading dashboard with signal generation, backtesting, portfolio tracking, and a Claude-powered trade journal. Kova is built for discretionary traders who want systematic signal support without giving up control.

Objectives

• Design a trading dashboard that surfaces AI signals alongside live portfolio positions
• Build a backtesting interface for validating strategies before deployment
• Create a Claude-powered trade journal for logging and reviewing decisions
• Deliver a dark-first, data-dense UI optimized for extended trading sessions

Approach & Activities

• Designed around the trader's core question: 'What should I be paying attention to right now?'
• Dashboard: real-time portfolio value area chart, active order panel, signal feed, positions table, and watchlist groups in one view
• Claude AI integration via dedicated API route for trade journal analysis and signal rationale
• Backtest page for running historical strategy simulations with Recharts visualization
• Dark-first design system with custom CSS variables for multi-monitor trading environments
• Full Next.js 16 + Prisma + Stripe stack with Clerk auth and SWR for data fetching

Challenges & Lessons

• Signal feed required clear BUY/SELL/WATCH distinction without overwhelming the dashboard with noise
• Backtesting UI had to feel responsive on large datasets without blocking the main thread
• Balancing AI-generated signal confidence with trader autonomy while ensuring the system advises and the trader decides

Impact