Skip to content
View Harlo9's full-sized avatar
🥶
Working from home
🥶
Working from home
3 followers · 6 following

Block or report Harlo9

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Content in all repositories owned by your account will be closed.
Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
Harlo9/README.md

Portfolio LinkedIn Email

I build RAG and multi-agent systems that run in production: retrieval, evaluation, deployment. Available for freelance missions.

Python LangChain LangGraph Ollama Azure OpenAI Anthropic FAISS FastAPI Docker PyTorch

🏭 Client work

Client What I built
Safran Aerosystems On-premise RAG assistant on their SharePoint (LangChain + Ollama). 250+ active users, reporting time cut by an estimated 60%.
Air Liquide R&D Japan RAG system for researchers: sourced answers, hallucination measurement, LLM-as-a-judge validation.

🧪 Projects

rag-benchmark-local-vs-cloud  RAG Evaluation Local vs Azure RAG benchmark on 6,000+ chunks of scientific papers. Dense, BM25, hybrid RRF and cross-encoder reranking compared on nDCG@10, then answer quality scored on faithfulness, hallucination rate and out-of-domain abstention.

NER_for_Cyber  NLP Cybersecurity Named entity recognition on cybersecurity text.

credit-scoring  ML End-to-end credit scoring model, from data prep to evaluation.

🌍 About

  • 🎓 MSc in Machine Learning & Deep Learning, IPSA Paris (2026)
  • 🇰🇷 6 months in South Korea (Keimyung University) · 🇯🇵 6 months in Japan (Tokyo)
  • 📱 Tech and AI content creator, 100K+ followers (@harlocmoi)
  • 🗣️ French (native) · English C1 (TOEIC 900/990) · Spanish B2
🟢 Open to freelance missions, remote or Paris.

Pinned Loading

  1. rag-benchmark-local-vs-cloud rag-benchmark-local-vs-cloud Public

    Local vs cloud RAG benchmark on a medical corpus. Cites its evidence, declines when the corpus cannot answer, and measures both.

    Python 3

  2. llm-credit-scoring llm-credit-scoring Public

    Local-first business loan analysis: an LLM extracts the data, deterministic Python computes the risk score. Ollama, Streamlit, 154 tests.

    Python 1

  3. agent-financier agent-financier Public

    Application macOS en barre de menu : un assistant financier connecté à l'API Anthropic, avec résumé hebdomadaire du portefeuille envoyé par email. Electron, node-cron, nodemailer.

    HTML 1

  4. cyber-ai-ner cyber-ai-ner Public

    Forked from laymiling/NER_for_Cyber

    Named Entity Recognition for Cybersecurity

    Jupyter Notebook 1

  5. image_recognition image_recognition Public

    We use a machine learning method, deep learning, to recognize various handwritten digits through a Pygame window

    Python