YNJ®
GitHub

AI & DATA SCIENCEPORTFOLIO / 2026

Yasaman
Najafi Jozani

Across applied AI, language technologies, and complex data, I like finding the pattern, testing what it means, and seeing where the idea leads.

Explore experiences GitHub profile
Yasaman Najafi Jozani standing beside a tree

01 / EXPERIENCE

Selected Experiences

Projects across data science and AI engineering.

AMS-02 detector mounted on the International Space Station AMS02 / DATA ANALYSIS THE ALPHA MAGNETIC SPECTROMETER ON THE ISS
DATA SCIENCE EXPERIENCE01 — 07

AMS-02: The Alpha Magnetic Spectrometer

At RWTH Aachen University, I led the end-to-end analysis of hundreds of billions of AMS-02 events, building reproducible data-integration, validation, and HPC workflows primarily in Python, with some C++ components. I used template fitting, Bayesian inference, and Monte Carlo simulations to measure photon flux and distinguish rare, weak signals from background.

  • Python
  • C / C++
  • Template fitting
  • Flux analysis
About AMS-02 View my project
DATA SCIENCE EXPERIENCE02 — 07

KiDS-1000 Survey

Feb 2020 – Apr 2021

For my master's thesis at the University of Bonn, I analysed approximately 21 million galaxy observations using Python (NumPy, SciPy, Pandas, and Matplotlib), building reproducible workflows for data validation and systematic-error correction. I applied Bayesian inference and likelihood-based models in CosmoSIS to estimate cosmological parameters, and correlation functions with E/B-mode decomposition to measure weak-lensing signals and control systematics. I also contributed statistical analysis and methodology validation to a peer-reviewed Astronomy & Astrophysics paper (2022).

  • Python
  • CosmoSIS
  • Bayesian inference
  • Weak lensing
KiDS-1000 weak-lensing data Read the publication
MULTIMODAL / VIDEO Q&A SEARCH · RETRIEVE · TIMESTAMP
AI ENGINEERING EXPERIENCE03 — 07

VideoQA: Ask YouTube Videos Anything

Built a multimodal AI chatbot that answers text or voice questions from YouTube videos. A LangChain agent finds relevant videos, embeds their transcripts in Pinecone, and answers with clickable timestamps, saying so when the videos don’t cover a question. Evaluated answer correctness and groundedness with LangSmith, and deployed the React frontend and FastAPI backend on Vercel.

  • LangChain
  • Pinecone
  • OpenAI API
  • FastAPI
  • React
Try the live app View source code
RAG / ENTERPRISE KNOWLEDGE RETRIEVE · ANSWER · CITE
AI ENGINEERING EXPERIENCE04 — 07

IronStore Enterprise AI Assistant

Built a RAG chat assistant that answers questions about IronStore’s internal policies and product catalogs and cites the source file and page for each answer. It retrieves the most relevant document chunks using OpenAI embeddings, restricts gpt-4o-mini to that context, and keeps recent chat history so follow-up questions work. Deployed the React frontend and FastAPI backend on Vercel.

  • RAG
  • OpenAI API
  • FastAPI
  • React
  • Vercel
Try the live app View source code
AI ENGINEERING EXPERIENCE06 — 07

Sentiment Analysis of Amazon Reviews

I collaborated on an NLP project analyzing Amazon customer reviews, combining TF-IDF and text embeddings with both fine-tuned pretrained Hugging Face models and our own model for sentiment classification. We also explored product-category clustering and concise review summaries, then deployed the interactive app on Streamlit.

  • NLP
  • Sentiment analysis
  • Clustering
  • Generative AI
Try the live app View source code
AI ENGINEERING EXPERIENCE07 — 07

AI Travel Advisor

Built a conversational travel advisor with the OpenAI API and a custom FastAPI backend. System instructions and tool integrations guide the model to return structured, day-by-day itineraries enriched with weather, currency, and transport information. Deployed the frontend and backend on Vercel.

  • OpenAI API
  • FastAPI
  • Python
  • Vercel
Try the live app View source code

02 / APPROACH

Curiosity is
a working method.

With a background in physics and more than five years of experience in big data analysis, I’m drawn to problems where careful thinking can reveal a useful signal. I bring rigorous analytical thinking and a strong grounding in statistics to complex questions, then carry that same curiosity into AI engineering—building projects that connect APIs and AI tools to practical applications.

01 / Applied machine learning02 / Language & generative AI03 / Experimental data analysis

03 / GET IN TOUCH

Have a good
question?

For collaborations, project conversations, or a thoughtful exchange.

yasaman.nj@gmail.com Find me on GitHub