Arjuna
An AI-driven autonomous trading engine for prediction markets focused on FIFA football matches. It uses a three-agent architecture to gather match data, generate predictions, execute orders on Polymarket and Kalshi, and learn from post-match outcomes through adaptive risk multipliers.
Project Highlights
Designed a three-agent architecture (Brain, Muscle, Memory) with decoupled subsystems communicating only through SQLite for resilience and independent scaling.
Built cross-exchange arbitrage detection using LLM-powered semantic matching to map equivalent markets between Polymarket and Kalshi.
Implemented synthetic stop-loss and take-profit brackets that monitor prices every 15 seconds since Polymarket lacks native bracket order support.
Created an adaptive risk system where daily failure analysis adjusts position sizing per strategy, market type, and regime based on real performance data.
Explore More Work
Deep dive into other high-performance solutions.
A browser-native collection of 120 plus developer utilities for formatting, debugging, conversion, and day-to-day engineering workflows. The product is optimized for privacy, fast local execution, offline use, and SEO-driven discovery.
A focused Mermaid diagram editor that turns raw Mermaid syntax into polished flowcharts and exports them as SVG, PNG, JPEG, or PDF. The app runs entirely in the browser, supports offline work, and keeps diagram data private on-device.
A privacy-first PDF studio for merging, splitting, OCR, converting, editing, and organizing PDF files directly in the browser. The product avoids uploads and accounts, giving users a local workflow for sensitive document handling.
Enterprise lead scoring built with XGBoost and Spring Boot to predict conversion likelihood and improve routing quality. The system improved conversion rates by 35 percent through better signal extraction and smarter recommendations.
A lightweight notification system that monitored appointment availability and sent location-aware alerts in real time. The product was designed for clarity, urgency, and speed during a high-demand public service moment.
A long-running technical blog covering Java, Spring Framework, and software engineering practices. It turns implementation experience into content that is useful for working developers and hiring teams evaluating technical depth.
An AI-powered stock research and automated trading platform for US equities and leveraged ETFs. It runs six LLM agent strategies that scan markets, analyze sentiment, score opportunities, and execute risk-gated trades through Alpaca in real time.
An AI-driven autonomous trading engine for prediction markets focused on FIFA football matches. It uses a three-agent architecture to gather match data, generate predictions, execute orders on Polymarket and Kalshi, and learn from post-match outcomes through adaptive risk multipliers.