"Turning Complex Data into Intelligent, Scaleable, and Actionable AI Solutions."
Welcome to my portfolio. I am an AI Manager & Developer specialized in bridging the gap between cutting-edge research and production-ready applications. This repository showcases my work in Generative AI, GraphRAG, MLOps, and Financial Engineering.
- The Problem: AI projects often remain as "local experiments" that fail to deliver real-world value due to high costs, poor accuracy (hallucinations), or lack of deployment discipline.
- The Solution: A multi-disciplinary skill set that combines Data Science, Symbolic Logic, and Full-stack Engineering to deploy robust, serverless AI agents.
- The Impact:
- 6+ Production Ready Projects: Deployed across Vercel and Streamlit Cloud.
- SOTA Integration: Leveraging Llama 3.3, Google Gemini, and GraphRAG architectures.
- Optimization Native: Reducing analysis time and operational overhead through intelligent automation and RAG (Retrieval-Augmented Generation).
| Domain | Tools & Technologies |
|---|---|
| AI / ML / GenAI | Python, TensorFlow, PyTorch, LangChain, LangGraph, OpenAI, RAG, Ollama |
| Data Engineering | Pandas, Polars, SQL, R, Google Cloud |
| Visualization | Plotly, Matplotlib, Seaborn |
| Fullstack & DevOps | React, Vite, Flask, Docker, Vercel, GitHub Actions |
| Vibe Coding | Antigravity, Cursor, Lovable, Gemini-Cli |
Automating Financial Intelligence
- The Problem: Retail investors lose hours manually calculating technical indicators.
- The Impact: Automates Fibonacci, MA, and S/R levels with 0 external TA libraries; provides AI-summarized insights in <500ms.
- Tech Stack: Python, Streamlit, yfinance, Plotly, Groq (Llama 3.3).
- 🔗 Code | 🌐 Live Demo
AI-Powered Editorial Strategy
- The Problem: High-quality SEO analysis is expensive and slow for content creators.
- The Impact: Generates structured reports with actionable keyword and readability insights instantly.
- Tech Stack: React, Vite, Groq AI, Vercel.
- 🔗 Code | 🌐 Live Demo
Accelerating QA Lifecycles
- The Problem: Writing BDD test cases manually is repetitive and error-prone.
- The Impact: Translates plain-text descriptions into standard Gherkin format using LLMs, reducing QA preparation time by 80%.
- Tech Stack: React, Vite, Google Gemini AI, Vercel.
- 🔗 Code | 🌐 Live Demo
Next-Gen Knowledge Retrieval
- The Problem: Traditional RAG ignores complex relationships between entities in unstructured data.
- The Impact: Explores the utility of Knowledge Graphs to provide more contextual and accurate RAG responses.
- Tech Stack: Python, Ollama, OpenAI.
- 🔗 Code
Low-Latency AI Communication
- The Impact: High-performance, embeddable multi-turn assistants and YouTube summarizers designed for speed and cost-efficiency.
- Tech Stack: React, Groq AI, YouTube API, Python.
- 🔗 Chatbot Repo | 🔗 Summarizer Repo
I am actively seeking opportunities to manage and build innovative AI solutions.
- 📧 Personal: marco.baturan@gmail.com
- 🏢 Professional: contact@getdevworks.com
- 💼 LinkedIn: in/marcogarciabaturan
- 📇 Virtual Card: hublink.getdevworks.com
Created with 💙 by Marco Baturan | Last Updated: 2026