🤝 Top Collaborators
.png&w=640&q=75)
🤓 Latest Submissions
TradOX
TradOX is an autonomous AI trading agent built on Alpaca's Trading API for the Alpaca AI Trading Agents Hackathon. It runs a complete decision loop: pull live and historical price data for a watchlist of large, liquid stocks (AAPL, MSFT, NVDA, GOOGL, TSM, AMZN), compute technical indicators, reason about them with an LLM, and execute real paper trades — all logged and explained. Rather than relying on a single noisy signal, TradOX combines three indicators — RSI, MACD, and SMA5/SMA20 trend — and only acts when at least two agree, reducing false signals from any one indicator alone. Every open position is protected by a hard-coded stop-loss that is completely independent of the AI's own judgment, so a bad model call can never go unchecked. Position sizing is cash-based rather than fixed, and the agent will never sell a stock it doesn't hold or stack duplicate orders on top of a pending one. The live dashboard shows real account P&L (both total return since baseline and today's change), current positions with live unrealized profit/loss, a full paginated decision log with the RSI/MACD/trend values behind every call, and manual controls: a pause/resume switch and a one-click liquidate button per position, so the human stays in control at all times. A "Goal of the Day" feature lets the user set a target return; once hit, the agent locks in gains and pauses automatically. TradOX is deployed live on Render with a background scheduler that keeps scanning and trading autonomously even when no one is watching — built with the busy, time-poor trader in mind, not someone glued to a screen all day. Planned next: multi-day/weekly goals with scheduled Discord check-ins, so the agent can be trusted to run unattended for longer stretches while still keeping the person genuinely informed, not just notified.
4 Sep 2026
TRIARC
TRIARC is a three-tier, token-efficient AI agent built for the AMD Developer Hackathon Track 1: Hybrid Token-Efficient Routing Agent. Rather than sending every task to the most powerful (and most expensive) model, TRIARC routes intelligently. A local Tier-1 router first classifies each incoming task by the capability it actually needs — extraction, classification, simple code generation, complex reasoning, debugging, research, or synthesis. The model registry then resolves that capability to the cheapest Fireworks-hosted endpoint that satisfies it: a mid-tier Gemma model for moderate tasks, or a frontier model reserved for genuinely hard ones. If a worker's response comes back with low confidence, TRIARC doesn't just accept it — it escalates the same task to a stronger, costlier endpoint and retries, rather than defaulting every task to the most expensive model out of caution. This keeps token spend proportional to actual task difficulty. The system runs as a self-contained batch container: it reads a set of tasks from /input/tasks.json, routes and executes each one through Fireworks AI, and writes structured results to /output/results.json — no server, no manual intervention, ready for automated evaluation. TRIARC also includes an optional local management API and terminal dashboard (Textual TUI) for monitoring runs, cost savings, and routing decisions during development — though the core submission scoring runs purely through the batch container against Fireworks AI.
13 Jul 2026
