Skip to content
View Nawfal-AI's full-sized avatar

Block or report Nawfal-AI

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
Nawfal-AI/README.md

Hi, I'm Nawfal 👋

MSc AI graduate (Università della Svizzera italiana, 2026) with a full-stack software engineering background. I build ML/NLP systems end-to-end — from raw data through to a working, documented, deployable system — and I'm currently looking for my first full-time role as a Data/ML Engineer and similar roles.

  • 🔭 Background in data pipelines, RAG, LLM fine-tuning, and MLOps (Docker, CI/CD, FastAPI)
  • 🌱 Currently finishing my thesis on few-shot visual learning for pest detection in precision agriculture
  • 💬 Open to Data/ML Engineer roles across Switzerland and the EU
  • 📫 Reach out: LinkedIn · nawfal1022@gmail.com

A few projects worth a closer look

Project What it shows
insurance_agent RAG + tool-calling agent (calculator, glossary lookup) over policy documents, served via FastAPI
EagleSFT-Bilingual-Dataset-and-LLM-FineTuning 536K-example bilingual (EN-RU) SFT dataset, built end-to-end, plus a Qwen2.5 LoRA fine-tune on top of it
DPO-Safety-Alignment-Phi3 DPO/QLoRA safety alignment of Phi-3-mini on a single T4 GPU, with real training runs and documented debugging
manipulation-policy-pipeline Imitation-learning policy with a full deployment pipeline — FastAPI, Docker, Kubernetes, Prometheus/Grafana
Action-Event-Focused-Captioning Team project comparing BLIP, ViT-GPT2, and GIT for action-focused video captioning, with full BLEU/METEOR/ROUGE evaluation
AI-Resume-Skill-Extractor NLP tool that scores resume-to-job-description match quality, packaged and containerized

Tools I work with

Python PyTorch LangChain Docker Kubernetes FastAPI SQL Git Hugging Face Transformers

Pinned Loading

  1. insurance_agent insurance_agent Public

    AI insurance coverage-check agent with a real tool-calling loop (RAG + calculator + glossary) over synthetic policy documents, served via FastAPI

    HTML 1

  2. EagleSFT-Bilingual-Dataset-and-LLM-FineTuning EagleSFT-Bilingual-Dataset-and-LLM-FineTuning Public

    EagleSFT: a 536K-example bilingual (RU/EN) dataset for LLM fine-tuning, plus the pipeline to build it and a Qwen2.5 LoRA fine-tune for category prediction and text generation. CC0 licensed.

    Jupyter Notebook 1

  3. Action-Event-Focused-Captioning Action-Event-Focused-Captioning Public

    Team project comparing BLIP, ViT-GPT2, and Microsoft GIT fine-tuned for action-focused image/video captioning. Includes BLEU/METEOR/ROUGE evaluation before and after fine-tuning, plus a video activ…

    Jupyter Notebook 1

  4. AI-Resume-Skill-Extractor AI-Resume-Skill-Extractor Public

    NLP tool that extracts technical skills from resume PDFs and scores match quality against a job description (keyword overlap + TF-IDF similarity). Packaged, tested, containerized, with a Streamlit …

    Python 1

  5. DPO-Safety-Alignment-Phi3 DPO-Safety-Alignment-Phi3 Public

    DPO fine-tuning of Phi-3-mini for safety alignment using QLoRA, runnable end-to-end on a single 16GB T4 GPU. Includes training health monitoring, data validation, and a toxicity/reasoning eval harn…

    Python 1

  6. manipulation-policy-pipeline manipulation-policy-pipeline Public

    Imitation-learning policy (BC and Diffusion Policy) for a manipulation task, with a full deployment pipeline. FastAPI serving, Docker, Kubernetes, Prometheus/Grafana monitoring, CI/CD.

    Python 1