Available for opportunities

I turn data into clear insights and scalable AI.

Data → |

I build data solutions from start to finish. My work covers everything from interactive Power BI dashboards that highlight business trends to machine learning models that solve real world problems. I focus heavily on clean code and measurable outcomes. Ultimately, I take messy datasets and turn them into clear strategies you can act on.

Rayan Farahani
Python
AI/ML
Analytics
15+ Projects Completed
5+ AI Models Deployed
10+ Dashboards Built
3+ Years Experience

Try My AI

Image captioning, hand-tracking drawing, and document Q&A. Built and deployed with no sign-up required.

Visual Assistance Image Captioning

FastAPI inference service. Upload an image to generate a natural language caption.

Model Details: Custom built architecture combining EfficientNetB0 for feature extraction and GPT-2 for text generation. Developed specifically to assist visually impaired individuals with daily life navigation.

Instructions: For testing, please upload images belonging to the following categories:

ATM · ATM queue · ATM use · Construction · bench · bus · chem · church · door · eurofifty · eurofive · euroten · eurotwenty · eurotwohundred · food_street · greensig · lift · luas · luas_pp · music · play · playing · poundfifty · poundfive · poundten · poundtwenty · push · redsig · stairsdown · stairsup · tactile · trash · wait · wallet · washroom · wetfloor

Finger Drawer

Draw in the air with your index finger using hand tracking. Ported from my Finger-drawer project and running directly in your browser.

How to Use Finger Drawer

  1. Open your hand: Keep your hand open with fingers extended
  2. Move your index finger: The tip of your index finger acts as the brush
  3. Close your fist: Make a fist to stop drawing and lift your hand
  4. Use controls: Change colors or adjust brush size
  5. Clear: Click "Clear" to erase your drawing

Camera access will be requested when you start. Make sure you are in a well-lit area.

DocMind AI

Upload up to 5 files (PDF, TXT, DOCX, CSV, XLSX, images) and chat with your documents. Powered by RAG, ChromaDB, and Mistral-7B, running live on Hugging Face.

No document loaded

Drop your files here

Up to 5 files · PDF, TXT, DOCX, CSV, XLSX, JPG/PNG/WEBP

or try a sample question
Powered by
LangChain ChromaDB Mistral-7B FastAPI HuggingFace
Open full app on HuggingFace
DocMind AI
Ready to analyze your documents
đź‘‹ Hi! I am DocMind AI, a RAG powered assistant built by Rayan.

Upload up to 5 files (PDF, TXT, DOCX, CSV, XLSX, images) on the left, then ask me anything. I will retrieve the most relevant sections and give you accurate, grounded answers.
DocMind AI · Just now
Answers are grounded in your documents only to prevent hallucinations.

Featured Projects

Multi-Task Real Estate Analytics

An AI system designed to accurately predict home prices and property types. By combining multiple models and cleaning complex data, this project provides highly reliable real estate market insights.

Pandas Plotly Streamlit Prophet

NSW Councils Financial Performance Dashboard

Interactive Power BI report analyzing revenue, expenses, and financial health across NSW local councils. Built to uncover spending patterns, funding structures, and long-term sustainability indicators.

Power BI DAX Data Modeling Financial Analysis

Customer Churn Analysis Dashboard

Built a churn analysis dashboard to identify at-risk customers, retention drivers, and revenue impact. Includes segmentation analysis, KPI tracking, and interactive filtering for decision makers.

Power BI DAX Data Cleaning KPI Design

Bank Marketing Analysis

Cleaned and prepared a bank marketing dataset, applied feature encoding, then trained and evaluated a Logistic Regression model. Reported results using a confusion matrix and achieved solid test accuracy.

Pandas NumPy Scikit-learn One-Hot Encoding Logistic Regression

Sea Turtle Image Segmentation

Designed and evaluated deep learning segmentation models for sea turtle detection. Compared baseline U-Net with attention enhanced U-Net + CBAM to measure performance improvements in pixel level classification.

PyTorch U-Net CBAM Computer Vision Segmentation

DocMind AI: Multi-Document RAG

Production ready RAG system that lets users upload multiple documents and ask grounded questions. It uses semantic search with MMR retrieval and a local Mistral 7B backend for accurate answers.

LangChain ChromaDB Mistral-7B FastAPI HuggingFace

Visual Assistance Image Captioning

Complete image captioning pipeline deployed as a FastAPI service on Hugging Face. It combines EfficientNetB0 for vision and GPT-2 for language to generate clear descriptions that help visually impaired individuals.

FastAPI PyTorch EfficientNetB0 GPT-2 HuggingFace

Review Similarity: Semantic Search

Semantic similarity search over e-commerce reviews using sentence transformers. Type any review and instantly retrieve the most similar feedback along with topic tags. Runs fully locally with no API keys required.

Sentence-Transformers Embeddings Cosine Similarity Gradio NLP

n8n RAG: Automated Support Workflow

Built an n8n based Retrieval Augmented Generation workflow for customer support. It ingests a knowledge base, stores vectors, retrieves relevant context, and drafts grounded replies. Deployed as a containerized Space for easy sharing.

n8n RAG pgvector Embeddings Automation

About Me

I am a Data Analyst and AI Engineer based in Sydney, Australia. I hold a Master's in IT with a focus on Artificial Intelligence.

I extract insights from complex datasets, build interactive dashboards, write optimized SQL queries, and deploy machine learning models that solve actual business problems.

I work with teams to translate data into business strategies that drive measurable results. My approach is simple: I combine analytical rigor with creative problem solving.

Location Sydney, Australia
Education Master of IT in AI
Status Open to opportunities
Core
Stack

Skills & Tools

Python

Data analysis, APIs, app deployment, and modeling

Pandas Pandas NumPy NumPy Scikit-learn Scikit-learn Streamlit Streamlit FastAPI FastAPI Gradio Gradio

SQL & Databases

Relational architecture, vector storage, and data pipelines

PostgreSQL PostgreSQL pgvector MySQL MySQL ChromaDB CTEs & Window Functions

Power BI

Interactive dashboards, data modeling, and business intelligence

Power BI DAX Power Query Data Modeling Dashboards

AI / ML

Deep learning, NLP, RAG, and Computer Vision

PyTorch PyTorch TensorFlow TensorFlow RAG & LLMs LangChain LangChain HuggingFace HuggingFace Computer Vision

Cloud & Automation

Deployment, infrastructure, and automated workflows

AWS Azure n8n Power Automate APIs & Webhooks

Let's Work Together

I am currently open to new opportunities. Whether you have a project in mind, a question, or just want to connect, my inbox is always open.