← Lavori

Python Developer — R/T Market Signal Engine (Futures / Databento)

Budget: $2250.0 FIXED / ⭐ 4.78 (127) United States

python, machine-learning, artificial-intelligence, data-science, python-sklearn

BUDGET: Fixed-price milestone: $2,250 Rate equivalent: $50/hr × 45 hours Payment: Upon verified acceptance — not on delivery ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ PROJECT OVERVIEW ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ We operate a proprietary automated trading system for CME futures instruments. The system is live in paper trading and moving toward a live account. The core infrastructure — signal processing, ML scoring, trade logging, and execution — is already built and operational. This project is a single focused deliverable: a Python service that generates trading signals in real time and delivers them to our existing backend via webhook. Think of it as replacing a third-party alert layer with a clean, internally-controlled Python service that we own and operate. Full technical details are provided under NDA. An NDA is required before scope documents, architecture diagrams, or system access are shared. ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ WHAT YOU'RE BUILDING ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ A Python service that: 1. Connects to Databento via WebSocket and streams real-time market data for 4 CME futures instruments 2. Constructs 2-minute OHLCV bars from the tick stream in real time 3. Evaluates two proprietary intraday strategies on each bar close: - An opening range breakout (ORB) strategy - A moving average bounce (MA Bounce) strategy Both strategies use standard technical indicators (SMAs, ATR, VWAP, ADX, volume). Logic is documented and will be provided under NDA. 4. When a signal condition is met, POSTs a JSON payload to our existing webhook endpoints (one per instrument). The payload format is fully documented — it must match our existing schema exactly. 5. Runs as a systemd service on a Linux VPS with a /health HTTP endpoint 6. Operates within defined trading session windows (configurable per instrument) The service must be clean, stable, and restartable. It does not place trades directly — it only generates and delivers signals to our system, which handles all execution logic. ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ WHAT ALREADY EXISTS (YOU ARE NOT STARTING FROM SCRATCH) ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ - Active Databento account with historical data already downloaded - 4 working webhook endpoints — tested, healthy, accepting JSON - Full strategy logic documented and available under NDA - Existing signal payload schema documented (your output must match it exactly) - VPS infrastructure already provisioned - A parallel validation framework: the new engine will run alongside the existing signal source for 2 weeks before cutover, allowing signal-by-signal comparison ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ TECHNICAL REQUIREMENTS ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ Required: - Python 3.11+ with asyncio / async/await - WebSocket client experience (streaming data, reconnection logic) - Real-time bar construction from tick or trade data - Technical indicator calculation (pandas-ta or equivalent) - JSON payload construction and HTTP POST delivery - systemd service deployment on Ubuntu Linux - Clean error handling, logging, and reconnection on disconnect Nice to have: - Prior Databento API experience - Experience with CME futures data feeds - Familiarity with intraday trading strategies (ORB, MA bounce) - FastAPI for the health endpoint ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ TIMELINE ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ Target: 3 weeks MAXIMUM Week 1-2: Build and validate against historical data Week 3-4: Parallel run alongside existing signal source; verify signal timing and accuracy; cutover This timeline is firm. We are actively trading and need this operational. ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ ACCEPTANCE CRITERIA ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ Payment is released upon verified acceptance — not on code delivery. Acceptance requires: - Service runs for 5 consecutive trading days post-cutover without manual intervention - Signal delivery confirmed in our server logs for all 4 instruments - /health endpoint returns correct status - Zero missed signals that the reference system would have generated If acceptance criteria are not met, contractor fixes at their own cost before milestone payment is released. ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ NDA REQUIREMENT ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ This project involves proprietary trading logic and infrastructure. An NDA is required before any technical details, architecture documents, or system access are shared. NDA will be provided at first contact. Do not apply if you are unwilling to sign an NDA. ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ TO APPLY ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ Include in your proposal: 1. Brief description of a WebSocket streaming project you have built 2. Your experience with real-time financial data or trading systems 3. Confirmation that you are willing to sign an NDA before scope review 4. Your availability to start within the next 7 days Fixed bid only. Do not apply if you need to negotiate the rate.
Apri su Upwork