AI & Data Science Student · Agentic AI Workflows · LLM Evaluation · Developer Tools
I’m an AI & Data Science student based in Bari, Italy, building practical AI systems, agentic workflows, and developer tools.
My focus is turning ideas into structured, usable software workflows: repository analysis, LLM-assisted development, prompt engineering, evaluation systems, local-first assistants, and reliable AI automation.
I’m early in my professional path, but I’m project-driven, fast at learning, and focused on building useful systems instead of chasing hype.
- LLM evaluation
- Agentic AI workflows
- AI-assisted software development
- Repository intelligence
- Local-first AI assistants
- Python automation
- Prompt engineering
- RAG and memory systems
- Developer tools
wikiHub is a repository-intelligence workflow that turns product ideas into structured engineering knowledge packs for AI-assisted development.
The goal is to help developers and AI coding agents understand strong open-source references before implementation.
Instead of asking an AI agent to build from vague instructions, wikiHub helps create a better technical context first.
Core ideas:
- GitHub repository research
- repository signal scoring
- architecture pattern extraction
- implementation reference analysis
- structured knowledge packs
- AI-assisted development preparation
Repository: wikiHub Showcase
J.A.R.V.I.S. is a local-first personal AI assistant built in Python.
It explores how a personal assistant can combine voice interaction, local LLM reasoning, persistent memory, RAG over documents, and lightweight PC automation.
Core ideas:
- local reasoning with Ollama
- Claude fallback for complex reasoning
- voice interface experiments
- persistent memory
- RAG over study documents
- Python-based automation
- modular assistant architecture
Repository: J.A.R.V.I.S.
I’m interested in AI systems that are practical, inspectable, and useful.
Some of the areas I’m currently exploring:
Idea → Repository Research → Knowledge Pack → AI-Assisted Build
User Context → Memory → Reasoning → Tools → Useful Action
I care about building AI workflows that improve clarity, reduce confusion, and help humans work better with software systems.
- Python
- JavaScript
- Markdown
- SQL basics
- Claude
- OpenAI models
- Ollama
- Local LLMs
- Prompt engineering
- LLM evaluation
- RAG pipelines
- ChromaDB
- Git
- GitHub
- FastAPI / Flask-style backends
- Automation scripts
- Data analysis basics
- Web development foundations
- AI-assisted coding workflows
I’m open to remote or hybrid opportunities in:
- LLM evaluation
- AI training
- data annotation
- prompt engineering
- junior AI development
- Python development
- chatbot testing
- AI-assisted software engineering
- Italian / English language evaluation for AI systems
I’m interested in building useful AI systems, not hype.
Good AI tools should:
- improve real workflows
- respect context
- be reliable enough to use
- stay understandable
- reduce friction
- help people build better things
- LinkedIn: Marco Antonio Carmine Abate
- GitHub: TacoengineerIT