Akrem D.

I build Python bots, automation and machine learning models that run on their own in production, and keep running long after I hand them over. One of my bots has been running for a client for 8 months, and one client has hired me five times. What I build Bots: Telegram, Discord and Reddit bots for alerts, monitoring, community management, access control and payments. They run unattended on a server, with logging, retries and alerts when something stops. Automation: Android emulator automation running more than 100 accounts, browser automation with Selenium and Playwright, and scraping with proxy rotation and session handling. Machine learning: around 35 prediction models for a sports analytics platform, from data collection and feature engineering to training, testing and deployment. I also turned a football prediction model into a full product with Stripe subscriptions, user accounts and a secure database. AI integration: a Telegram bot that combines scraping with OpenAI to deliver player stats, and a hybrid search engine over 194,000 job listings with LLM reranking. Why clients keep working with me Most bot developers can't build a model, and most data scientists don't ship a bot that stays alive for months. I do both, and I deploy what I build, so you get a system that runs, not a script that only works on my machine. How I work Clear scope, milestones and honest timelines. If something looks risky or will take longer, I tell you early. You get working code, deployment and documentation. Master's in Data Science (highest honors). Arabic, English and French. Based in UTC+1 with no daylight saving, so the overlap with Europe never changes. Stack: Python, FastAPI, Flask, PostgreSQL, MongoDB, Redis, Selenium, Playwright, scikit-learn, XGBoost, LightGBM, PyTorch, OpenAI and Claude APIs, Stripe, Docker, Linux. Tell me what you want to automate, and I’ll come back with a plan and a timeline.