Hello, I'm

Uzma Khatun

AI/ML Engineer | Data Scientist

I'm an aspiring AI/ML Engineer with hands-on experience in building intelligent applications using Python, AI Framwork, and No-Code Automation tools. I love working with real-world datasets and deploying interactive web apps.

Explore More Hire Me Resume
Uzma Khatun

About Me

I'm a BCA student and self-taught ML developer passionate about building data-driven applications. My journey in AI/ML began with a curiosity about how machines can learn from data, and it has grown into a full-fledged passion.

I enjoy solving real-world problems using data and building interactive ML-powered applications. My journey in AI/ML has led me to work on several practical case studies and projects that demonstrate my skills in data analysis, machine learning, and deploying user-friendly web apps. I aim to contribute meaningfully to the tech community while continuously learning and growing.

When I'm not coding, you can find me reading research papers, participating in Kaggle competitions, or contributing to open-source projects in the AI community.

Skills & Technologies

Python

Advanced

Data Analysis

Advanced

Machine Learning

Advanced

Deep Learning

Intermediate

NLP

Intermediate

n8n

Intermediate

LangChain

Intermediate

CrewAI

Intermediate

Generative AI

Intermediate

Databases

Intermediate

Streamlit

Advanced

Version Control

Intermediate

Experience

April 2025 – September 2025

AI/ML Engineer Intern

AffyCloud IT Solutions

  • Built an AI Decision Support Agent that recommends whether to buy, sell, or hold company stocks.
  • Designed and implemented the system using LangChain initially, later enhanced with CrewAI for multi-agent decision-making.
  • Integrated real-time stock data and analytical logic to assist users in investment decisions.
  • Focused on improving accuracy, interpretability, and automation of financial insights.
LangChain CrewAI Agentic AI Python yfinance
August 2025 – September 2025

AI Engineer Intern

BXB Analytics

  • Developed an AI-powered Bedtime story generator and illustration system using no-code automation (n8n).
  • Integrated multiple AI agents, Perspective AI for toxicity and child-safety checks, and image generation models for visuals.
  • Used Supabase for user logs and cloud storage for generated images.
  • Connected workflows to a frontend (HTML, CSS, JS) via webhooks, allowing users to download storybooks as PDFs.
n8n Supabase Perspective AI Cloud Storage Webhooks
September 2025 – September 2025

Programmer Analyst Intern

Afucent

  • Built and automated Email Automation and LinkedIn Post Generation workflows using n8n.
  • Implemented custom nodes and API integrations to streamline repetitive business operations.
  • Enhanced workflow reliability and ensuring smoother automation performance.
  • Strengthened understanding of workflow orchestration, no-code platforms, and automation strategies for business efficiency.
n8n Workflow Automation API Integration

Featured Projects

Decision Support

Decision Support Agent

An AI-powered system that combines multiple specialized agents to analyze stock data, company fundamentals, and market trends. Providing smart Buy, Sell or Hold recommendations to support data-informed investment decisions.

Code Coming Soon
FableBug AI

FableBug AI

Developed an intelligent bedtime story generator using multi-agent AI workflow with n8n (no-code automation). Integrated OpenAI for story generation, Perspective API for content safety, and Pollination AI for illustrations.

Code Demo
PDF Que-Ans Bot

PDF Que-Ans Bot

Built a question-answering system that extracts answers from uploaded PDF documents using Groq LLM and SentenceTransformers. Developed with Streamlit and integrated semantic search for accurate context retrieval.

Code Demo
LinkedIn Corousel Post

LinkedIn Carousel Posts

Automated LinkedIn carousel generator using n8n, OpenRouter, and LinkedIn API. Scrapes 10 RSS feeds, filters articles by topic/date, generates carousel content, designs branded slides, and publishes directly to LinkedIn.

Code Demo
AI Chatbot

AI Chatbot with LangGraph

Developed an AI-powered chatbot using Groq LLM and Tavily API for real-time web search. Backend built with FastAPI and deployed on Render, frontend on Hugging Face Spaces via Streamlit.

Code Demo
Story Generator

AI Story Generator

Created an AI-powered tool that generates creative stories based on user-selected genre, theme, and characters. Built using Groq LLM and Streamlit for an interactive storytelling experience.

Code Demo
Adaptive Learning

Rule-Based Adaptive Learning

A rule-based math adaptive learning system that adjusts question difficulty in real time based on the learner’s accuracy and speed. It uses predefined logic to shift users between easy, medium, and hard levels.

Code Demo
PDF to Excel Converter

PDF to Excel Converter

AI-powered Python application that intelligently converts unstructured PDF data into structured Excel spreadsheets using Groq LLM. Eliminating manual data entry and enabling instant document analysis.

Code Demo
Car Price

Car Price Prediction

A machine learning model that predicts car prices based on features like brand, model, year, fuel type, and mileage using regression algorithms and data preprocessing techniques.

Code Demo
Student Performance

Student Performance Analyzer

A data-driven tool that analyzes student academic performance using machine learning classification models. It identifies key factors affecting student outcomes and helps predict performance based on features like study time, sleep hours, and past scores.

Code Demo

Case Studies

Airbnb Analysis

Airbnb Listing Analysis

Performed end-to-end data analysis on Airbnb listing dataset. Implemented data preprocessing pipeline. Built predictive models using Random Forest to forecast listing prices based on location, ratings, and property characteristics.

Code
Loan Approval

Loan Approval

Used Logistic Regression to predict loan approvals based on income, credit history, and employment status. Included data preprocessing, feature engineering, and model evaluation to support decision-making.

Code
Porter Time Estimate

Porter Time Estimate

Built a neural network to predict delivery times using features like vehicle type, distance, and traffic. Focused on data cleaning, feature selection, and model tuning to improve accuracy and efficiency.

Code
Patient Segmentation

Patient Segmentation

Used clustering (K-Means, TSNE) to group patients by age, BMI, blood pressure, and health conditions, enabling targeted healthcare strategies and personalized treatment plans.

Code
Super Store case study

Super Store Case Study

Analyzed sales data to uncover trends in profit, region-wise performance, shipping delays, and top-selling products. Delivered insights using Python and data visualization tools to improve business decisions.

Code
Crypto Sentiment Trader PnL Analysis

Crypto Sentiment vs PnL

A study combining Bitcoin sentiment data with trader PnL to analyze predictive patterns. Results show an inverse trend—higher “Greed” often precedes lower trader profitability.

Code

Get In Touch

I'm always excited to collaborate on innovative AI/ML projects and discuss opportunities in data science. Whether you have a project in mind or just want to connect, I'd love to hear from you!

uzmakhatun0205@gmail.com