Welcome to My Portfolio!
I’m a final-year Computer Science student passionate about solving data problems using Python, SQL, and MLOPS. 📈
I’ve built several end-to-end projects and aspire to work in data analytics and AI.🔍
About Me
I’m a Computer Science (Data Science) undergraduate from NIET, with a strong passion for uncovering insights from data to solve real-world problems.
My journey began with a curiosity for numbers and patterns, which evolved into hands-on experience through academic projects and self-driven learning.
I’ve successfully completed several data analysis projects that reflect my analytical thinking, technical skills, and commitment to continuous growth.
I enjoy transforming raw data into actionable insights. Explore my portfolio to see how I solve complex data challenges with practical, results-oriented solutions.
Skills & Technologies
Programming Languages
Python
SQL
Java
Data Visualization
Power BI
Tableau
Matplotlib / Seaborn
Databases & Tools
MySQL
MongoDB
Analytics & ML
Machine Learning
Statistical Analysis
Data Mining
Business Tools
Advanced Excel
Google Analytics
Cloud & Other
AWS S3
Git / GitHub
Kubernetes
Featured Projects
A robust end-to-end MLOps pipeline designed for managing vehicle insurance data — covering data ingestion, transformation, model training, and deployment with CI/CD automation. This project demonstrates real-world data management and deployment practices integrating AWS, Docker, and GitHub Actions.
Key Achievements:
- Developed automated data ingestion from MongoDB to ML pipeline
- Implemented data validation, transformation, and model training modules
- Deployed ML model using AWS S3, EC2, ECR, and GitHub Actions
- Enabled CI/CD workflow with Docker integration for seamless deployment
Built an interactive news research assistant using Google Gemini and Streamlit. RockyBot allows users to extract, analyze, and query multiple news articles intelligently using LLM-driven question-answering.
Key Features:
- Multi-article processing and content extraction via scraping and LangChain loaders
- Semantic search using vector embeddings (FAISS)
- Natural language Q&A with contextual memory and random prompts
- Visual metrics dashboard and quick insights with chat history
Comprehensive analysis of the Atliq Hardware Sales , focusing on market penetration, growth opportunities, and competitive landscape for strategic decision-making.
Key Achievements:
- Identify high and low-performing cities based on profit margin
- Track revenue and profit changes from 2017 to 2020
- View revenue and profit contribution by market
- Analyze revenue and profitability by customer.
An AI-powered research assistant built with Streamlit, LangChain, and Groq (LLaMA3) — designed to summarize uploaded PDFs/TXT, answer document-based questions with source references, generate logic-based MCQs, and run conversational Q&A backed by FAISS vector search.
Key Features:
- Upload and process .pdf / .txt research documents
- AI-generated concise summaries with source citations
- Document Q&A with conversational memory buffer
- "Challenge Me" mode: logic-based MCQs + feedback
- Fast inference using Groq LLaMA3; semantic search via FAISS
- Secure API key handling (.env or Streamlit secrets)
Let's Connect!
I'm always excited to discuss data projects and opportunities