class AmeerHamza:
location = "Istanbul, Turkey πΉπ·"
education = "MS Artificial Intelligence β Beykoz University"
focus = ["LLM Systems", "RAG Pipelines", "Multimodal AI", "Data Engineering"]
deployed = ["sktech-ai.vercel.app"]
published = ["DOI: 10.5281/zenodo.15856940"]
open_to = "Remote AI/ML Engineering and Data Engineering roles"I build production-grade AI systems and data pipelines β from multimodal RAG tools that convert whiteboard sketches to diagram code, to Databricks medallion pipelines processing real e-commerce data at scale.
| Project | Description | Stack | Status |
|---|---|---|---|
| β‘ SketchFlow AI | Converts whiteboard sketches and text to Mermaid/PlantUML diagram code using Gemini Vision + ChromaDB RAG + multi-provider fallback chain | Gemini Β· Groq Β· OpenRouter Β· ChromaDB Β· FastAPI Β· React | π΄ Live |
| π§ DocuMind AI | Multi-source RAG assistant ingesting PDFs, web URLs, and YouTube transcripts with streaming citation-backed answers | FastAPI Β· Groq Β· ChromaDB Β· Sentence-Transformers Β· React | β Complete |
| Project | Description | Stack | Status |
|---|---|---|---|
| π OrderStream Pipeline | End-to-end e-commerce pipeline on Databricks β Bronze/Silver/Gold medallion architecture, PySpark, Delta Lake, 4-task Job orchestration, AI/BI dashboard | Databricks Β· PySpark Β· Delta Lake Β· Unity Catalog | β Complete |
| π‘ SysStream | Real-time system metrics streamed via Kafka into SQL Server with live Power BI dashboards | Kafka Β· SQL Server Β· Python Β· Power BI | β Complete |
| π€ Airflow Weather ETL | Automated ELT pipeline β Open-Meteo API β Airflow DAGs β dbt models β PostgreSQL, Dockerized | Airflow Β· dbt Β· PostgreSQL Β· Docker | β Complete |
AI / LLM Engineering
Data Engineering
Frontend & Deployment
- π Loop Engineering: Why I Stopped Prompting My RAG Pipelines and Started Designing Loops
- π From Simple RAG to Agentic RAG: When, Why, and How to Use Them
- π Multimodal AI in 2025: Why Combining Text, Image, Audio & Video Is the Future
- π Data Quality for Real-World AI: Enforce it with Pytest & Great Expectations
NexGen MediaCraft: Real-Time AI News Anchoring with LLM-Based Generation and Web Scraping
Ameer Hamza, Nimra Mairaj Β· Zenodo, July 2025
- Extending OrderStream with Databricks Auto Loader and Delta Live Tables
- Writing about loop engineering patterns for production LLM systems
- Open to remote Junior AI/ML Engineering and Data Engineering roles
"The goal is not to build AI that replaces humans but AI that makes humans more capable."
β If you find my work useful, a star means a lot!