🔭 I’m Amin Zay - aFull Stack Data Scientist / Senior AI Engineer working at the intersection of real-world data, scalable backend systems, and modern LLM architectures.

💡 I believe the real challenge in LLM adoption isn’t the model, it’s the pipeline.

Over the past decade, I’ve evolved from building classical ML models to designing robust data and AI systems that embed LLMs into business logic, not just the UI. My focus is onLLM pipelines, context-aware agents, andscalable AI infrastructure.

🔹 I recently led the integration of ChatGPT APIs with Retrieval-Augmented Generation (RAG) in a high-availability backend, where real-time inference demanded smart data routing and prompt engineering.

🔹 In healthcare, I’ve driven the development of LLM-based analytics tools using domain-tuned prompting and fine-tuning to improve patient outcomes.

🔹 My engineering toolkit includes: Django, Redis, PostgreSQL, Docker, AWS, CI/CD pipelines - all optimized for production-level AI workloads.

🧠 I’m currently exploring Agent Communication Protocol (ACP) and Model Context Protocol (MCP), where LLMs evolve into decision-making collaborators across pipelines.

What excites me most:

• ✅ AI agents embedded in developer workflows & analytics platforms
• ✅ Semantic search & contextual reasoning powered by real-time pipelines
• ✅ Architecting ML systems thatdeliver, not just demonstrate

Let’s connect if you’re working on:#LLM #RAG #AIEngineering #PromptEngineering #MCP #ACP #SemanticSearch #ScalableAI #DataPipelines #MLinProduction #RealWorldAI

• Expert: Python, SQL, JavaScript, PHP, C++, C#, VB
• Familiar: R, MATLAB, Java

• Back-End: Django, Django REST Framework, FastAPI, Flask, Redis, PostgreSQL
• Data Engineering: ETL Pipelines, Airflow (if used), Kafka, Elasticsearch
• DevOps: Docker Compose, Nginx, CI/CD, Git, AWS (EC2, Lambda, Sagemaker, Textract), GCP
• Monitoring & Deployment: Linux (CentOS), High Availability Systems (99.9% uptime)

• Core Skills: NLP, Machine Learning, Deep Learning, Prompt Engineering
• LLMs: ChatGPT API, RAG Architectures, LangChain, Transformers (HuggingFace), Chain-of-Thought, Few-shot/Zero-shot prompting
• Libraries: Pandas, NumPy, Scikit-learn, Matplotlib, TensorFlow, Keras, PyTorch

• Microservices Architecture, Domain-Driven Design (DDD), Data-Driven Design
• Design Patterns: MVC, TMP, Decorator
• Semantic Search, Real-time Inference Systems

• HTML, CSS, Bootstrap, JavaScript (jQuery), D3.js
• Interactive Dashboards & Data Visualizations (Tableau, Custom JS)

• Agile, Scrum, Kanban, Scrumban
• Project Management: Jira, Trello
• Cross-functional Team Leadership & Mentoring

• Amin University Data Science
• Deep Learning
• Data Mining
• K-Means
• DBSCAN Clustering
• KNN Regressor Evaluation
• Hierarchical Clustering
• Iran Stock Prediction
• University Projects

• My Python Code
• Python Django
• Python Flask
• My PHP Code
• PHP Simple Cache Engine
• C-Apps
• CPPApps

PinnedLoading

1. Amin-University-Data-ScienceAmin-University-Data-SciencePublic

   All data mining machine learning algorithms are basically coded by displaying solutions with Python

   Jupyter Notebook 5

2. PythonCodeSnippetsPythonCodeSnippetsPublic

   A collection of commonly used Python code snippets for various programming tasks.

   Python 9 3

3. My-Python-CodeMy-Python-CodePublic

   My Algorithms & coding with Python

   Python 4 1

4. Iran-Stock-PredictionIran-Stock-PredictionPublic

   Iran Stocks Prediction With LSTM Model from Keras Lib

   Jupyter Notebook 2 1

5. Docker-ProjectsDocker-ProjectsPublic

   Python 2 1

6. CPPAppsCPPAppsPublic

   All Cpp Apps (Algorithms)

   C++ 3