{
"developer": "Vasudev Talaviya",
"education": "B.C.A. in AI & Data Science",
"location": "Surat, Gujarat, India 🇮🇳",
"currently_building": "Personal-book-RAG (semantic search with exact citations)",
"core_focus": ["Data Analysis", "Applied ML", "Interactive Streamlit Dashboards"],
"philosophy": "Turning raw data and models into intuitive, living applications"
}| Project | What it is | |
|---|---|---|
| 📚 | Personal-book-RAG 🆕 Latest |
Ask a question in plain language and get the exact page from my own books and notes. |
| 🔍 | Smart-Detection | A trained computer vision model designed for object detection. |
| 📊 | Gen-Z-analysis | A study of 1,000,000+ records to understand the habits and choices of Gen Z. |
| 📶 | Airtel-Case-Study 💼 Case Study |
Telecommunications analytics evaluating market trends, user growth, and business patterns. |
| 🌱 | Green-Initiatives-of-Google 🌱 Case Study |
Sustainability analytics analyzing Google's carbon neutrality, clean energy, and green operations. |
| 🍎 | FitGenZ | A health platform that turns a photo of your meal into simple, useful health insights. |
| 🌍 | Life-expectancy-who | A study of global health data to see what affects how long people live. |
| 🎵 | my_music_enjoy | A personal project built around music and audio exploration. |
| 🌐 | Django | My practice space for learning backend architecture and web development. |
| 🔢 | Number-System- | A foundational project on how number systems work and convert. |
👉 See all 18 repositories on my profile →
|
Python ChromaDB Transformers BM25 Ask plain-language questions and retrieve exact passages with page-level citations from personal books and notes. 
|
Python Computer Vision Machine Learning A computer vision project focused on automated detection, image processing, and model classification. 
|
|
Python Pandas NumPy Streamlit In-depth exploratory data analysis and interactive dashboard examining 1,000,000+ demographic records.
|
🍎 FitGenZPython Data Analysis Health Tech A health platform that transforms food photos and dietary data into actionable wellness insights.
|
|
Business Analytics Python Telecom EDA Telecommunications case study examining telecom subscriber distributions, market penetration, and business metrics.
|
Sustainability Data Analytics Python Visualization Data-backed case study investigating Google's global sustainability reports, carbon footprint metrics, and green energy investments.
|
Ask a question in plain language. Get the exact page from my own books.
Instead of flipping through hundreds of pages, this project reads my PDF books and notes, understands them, and finds the passages that best answer a question, along with the book and page they came from.
flowchart LR
A[📄 Books & Notes] --> B[✂️ Clean & Split]
B --> C[🧠 Understand Meaning]
C --> D[(🗄️ Local Database)]
Q[❓ My Question] --> E[🔎 Meaning Search]
Q --> F[🔤 Keyword Search]
D --> E
D --> F
E --> G[⚖️ Combine Results]
F --> G
G --> H[🏆 Re-rank the Best]
H --> I[✅ Top Passages + Page]
classDef cyan fill:#161b22,stroke:#00D4FF,stroke-width:2px,color:#ffffff;
classDef purple fill:#161b22,stroke:#7C3AED,stroke-width:2px,color:#ffffff;
class A,B,C,D cyan;
class Q,E,F,G,H,I purple;
💡 What makes it special
| 🧩 Smart splitting | Books are cut into meaningful pieces, not random chunks, and every piece remembers its book and page. |
| 🤝 Two kinds of search | Meaning-based search catches ideas. Keyword search catches exact terms. Together they miss less. |
| 🏆 Re-ranking | A second pass reads the question and each result together to put the best answer first. |
| 🔬 Honest evaluation | I wrote a deep analysis of the first version's weak points, then a plan to fix them, and an evaluation script to measure the results. |
Tech Used: Python • PyMuPDF • ChromaDB • Sentence Transformers • BM25 • Cross-Encoder
Computer vision pipeline focused on automated object detection.
An end-to-end computer vision implementation designed to take image inputs, run model inference, and output bounding boxes with detection classifications.
flowchart LR
A[📷 Input Image] --> B[⚙️ Preprocessing & Resizing]
B --> C[🧠 Neural Model Inference]
C --> D[🎯 Bounding Boxes & Classes]
classDef cyan fill:#161b22,stroke:#00D4FF,stroke-width:2px,color:#ffffff;
classDef purple fill:#161b22,stroke:#7C3AED,stroke-width:2px,color:#ffffff;
class A,B cyan;
class C,D purple;
One million records. One question: what is Gen Z really like?
- Studied career choices, mental health patterns, and everyday habits across a dataset of 1,000,000+ people.
- Cleaned the data, found the patterns, and turned them into clear visuals.
- Packaged everything into an interactive dashboard so anyone can explore the data themselves.
Tech Used: Python • Pandas • NumPy • Streamlit • Plotly • Seaborn
flowchart TB
ME((🧑💻 Vasudev))
ME --> D[📊 Data Analysis]
ME --> M[🤖 Machine Learning]
ME --> W[🌐 Web & Backend]
ME --> S[🔎 Smart Search]
D --> D1[Life Expectancy<br/>WHO data]
D --> D2[Gen-Z Analysis<br/>1M records]
D --> D3[Airtel Study<br/>telecom trends]
D --> D4[Google Green<br/>sustainability]
M --> M1[Smart-Detection]
M --> M2[FitGenZ<br/>meal photo insights]
W --> W1[Django practice]
W --> W2[My Music Enjoy]
S --> S1[Personal-book-RAG]
classDef root fill:#161b22,stroke:#7C3AED,stroke-width:2.5px,color:#ffffff;
classDef branch fill:#161b22,stroke:#00D4FF,stroke-width:1.5px,color:#ffffff;
class ME root;
class D,M,W,S,D1,D2,D3,D4,M1,M2,W1,W2,S1 branch;
Click to expand / collapse
Dev Reality:
Q: Why do programmers prefer dark mode and headphones?
A: Because light attracts bugs, and noise ruins the vibe. 🎧✨