TakeProfit.am — Quantitative Algorithmic Trading Terminal
Institutional-grade systematic trading platform featuring real-time pattern recognition, multi-tenant proxy isolation, and low-latency Binance execution.
TakeProfit.am is a high-performance quantitative trading terminal engineered for algorithmic crypto trading. The platform scans dozens of candlestick patterns and technical indicators in real time across hundreds of pairs, providing an advanced backtesting suite, walk-forward analytics, and a dual-engine execution pipeline for paper and live trading on Binance Futures.
Technology stack
Problem
Algorithmic traders face severe execution bottlenecks, API rate-limiting issues, and static IP whitelisting constraints when running automated trading loops. High-frequency pattern evaluation also tends to flood users with duplicate, noisy signals across overlapping timeframes, leading to poor risk management.
Solution
Architected a global-calculation, local-filtering asynchronous engine that evaluates market patterns centrally and filters signals per user in under 15ms. Implemented a dynamic per-user outbound proxy binding system to assign dedicated static IPs to each client, alongside a strict timeframe-hierarchy deduplication algorithm that suppresses lower-timeframe noise.
Technical contribution
Designed and built the full-stack system from scratch: the Django/Celery asynchronous backend engine, dynamic WebSocket data streams, the Next.js responsive web app, a desktop app wrapper in Photino.NET (C#), a custom multi-tenant proxy router, and the Dockerized Hetzner infrastructure.
Architecture
Distributed hybrid architecture: centralized Celery workers aggregate live Binance data into Redis/PostgreSQL, while user execution tasks dynamically bind to isolated datacenter proxies (IP:PORT:USER:PASS) to preserve dedicated IP whitelisting. The Next.js dashboard is bundled into a cross-platform Photino.NET (C#) desktop client with automated startup cache-purging hooks.
Performance considerations
Sub-15ms database query response using composite indexes across candle boundaries. Timeframe conflict resolution runs in linear time to eliminate redundant signal noise, and a multi-queue Celery topology isolates global data scraping from high-priority order routing.
Security considerations
AES-128 API key encryption, strict outbound static IP proxy binding satisfying Binance's trusted IP restrictions, mandatory zero-withdrawal permission enforcement, and isolated multi-tenant execution pools.
Deployment model
Containerized multi-service setup with Docker Compose hosted on Hetzner dedicated CPU servers, with an automated macOS DMG distribution pipeline including a drag-and-drop installer layout and Apple code-signing.
Outcome
Successfully deployed a high-scale institutional trading terminal capable of executing hundreds of strategy loops concurrently with isolated per-user outbound IPs, zero execution collisions, and comprehensive risk-management controls.