I am a Professional Python Developer with over 20 years of experience in the technology ecosystem. My expertise spans four main areas of Python:

I believe in writing clean, documented, and testable code following PEP8 standards, using Type Hints and Design Patterns in Python projects.

Problem Solved: With every git push on the main branch, the Python project is automatically tested, built, and deployed on VPS with Docker. Reducing deployment time from 30 minutes to 2 minutes.

Key Technologies:

• Docker & Docker Compose
• GitHub Actions / GitLab CI
• Supervisor / systemd
• GitHub Secrets

Technical Highlights:

• Multi-stage Dockerfile for 70% image size reduction
• Smart entrypoint.sh script for automatic database migrations
• Secure token management with GitHub Secrets
• Auto-restart with Supervisor
• Telegram deployment notifications

Problem Solved: Combining classical econometric models with deep learning to predict gold, dollar, and oil prices with higher accuracy than traditional models.

Key Technologies:

• Statsmodels (VAR, ARIMA)
• Arch (GARCH)
• PyTorch / TensorFlow (LSTM, Transformer)
• yfinance / Alpha Vantage API
• Prophet (Facebook)
• Ray (Hyperparameter Optimization)
• XGBoost (Market Regime Detection)

Technical Highlights:

• Backtesting Engine with transaction costs and slippage
• Market regime detection (bullish, bearish, neutral) with XGBoost
• Ensemble approach with different models per regime
• Streamlit dashboard for prediction results
• Historical data storage in InfluxDB

Problem Solved: Automated receiving and responding from 6 platforms (WhatsApp, Telegram, Instagram, Rubika, Bale, Eitaa) using Generative AI (RAG) with conversation history. Reducing response time from 10 minutes to 10 seconds.

Key Technologies:

• Telethon (Telegram)
• instagrapi (Instagram)
• pywhatkit / selenium (WhatsApp Web)
• Rubika API / Bale API / Eitaa API
• LangChain (RAG)
• ChromaDB (Vector Database)
• HuggingFace Transformers (Llama/Mistral)
• FastAPI

Technical Highlights:

• Adapter Pattern for unified connection to all messengers
• Conversation Memory for context-aware responses
• Local models instead of cloud APIs for data privacy
• Persian language detection and timely responses
• File, image, and inline keyboard support

Problem Solved: Automatic retrieval of advanced metrics (engagement rate, reach, follower growth rate, peak activity hours) and actionable recommendations to improve content and increase engagement by up to 40%.

Key Technologies:

• Instagrapi
• Pandas / Polars
• Plotly / Streamlit
• Prophet
• HuggingFace Transformers
• ReportLab

Technical Highlights:

• Sentiment Analysis of comments and DMs with Persian language models
• Best posting time detection based on historical engagement patterns
• Automated weekly PDF report generation and email delivery
• Influencer identification
• Engagement graph visualization

Problem Solved: Direct connection to WooCommerce database and REST API, extracting order, product, and customer data, displaying key KPIs (AOV, Conversion Rate), best-selling products, and inventory status in real-time.

Key Technologies:

• WooCommerce REST API
• SQLAlchemy
• Pandas
• Streamlit / Dash
• APScheduler
• Plotly
• Prophet

Technical Highlights:

• Abstraction layer for data from two sources (API and direct database)
• Net profit calculation with product cost deduction
• Monthly, daily, and hourly sales trend charts
• Sales forecasting with Prophet
• Inventory alert system
• Excel/CSV report export
• Star Schema data modeling

Problem Solved: Automated scanning of newly listed coins on exchanges (Binance, Dexscreener, etc.) with multi-factor analysis (technical, sentiment, social networks) to identify coins with at least 50% growth potential in 7 days.

Key Technologies:

• ccxt
• DexScreener API / CoinGecko API
• Tweepy
• BeautifulSoup / Scrapy
• scikit-learn / XGBoost
• HuggingFace Transformers
• TA-Lib
• Redis
• Streamlit
• APScheduler

Technical Highlights:

• ETL data pipeline running every 5 minutes
• Classification model trained on historical data
• Scoring system based on:
  • Liquidity
  • Holders count
  • Twitter sentiment analysis
  • Technical indicators (RSI, MACD, volume)
• Suspicious pump detection and filtering
• Prioritized results with "Deep Analysis" button
• Comprehensive success factor reports
• Telegram alert system for high-score coins