import gradio as gr
import base64
import os
# --- Helper Functions ---
def file_to_data_uri(filepath, mime_type="application/pdf"):
with open(filepath, "rb") as f:
data = f.read()
b64 = base64.b64encode(data).decode("utf-8")
return f"data:{mime_type};base64,{b64}"
# --- Navigation Functions ---
def show_data_analytics():
return gr.update(visible=False), gr.update(visible=True), gr.update(visible=False), gr.update(visible=False)
def show_machine_learning():
return gr.update(visible=False), gr.update(visible=False), gr.update(visible=True), gr.update(visible=False)
def show_computer_vision():
return gr.update(visible=False), gr.update(visible=False), gr.update(visible=False), gr.update(visible=True)
def go_home():
return gr.update(visible=True), gr.update(visible=False), gr.update(visible=False), gr.update(visible=False)
# --- Icons (SVG) ---
data_analytics_icon = """Data Analytics
I specialize in transforming raw data into actionable business insights that drive strategic decision-making.
With a strong background in both data analytics and commerce, I bridge the gap between business needs and technical solutions.
My approach combines statistical analysis with compelling data visualization to tell stories that solve real-world problems.
I've developed expertise in designing dashboards that make complex data accessible and creating end-to-end analysis
workflows that uncover hidden patterns and trends.
""")
gr.HTML("""
Skills
Data Visualization
Power BI
Tableau
Matplotlib/Seaborn
Plotly/Dash
Data Manipulation
SQL
Pandas
NumPy
ETL Pipelines
Analysis Techniques
Statistical Analysis
A/B Testing
Time Series Analysis
Customer Segmentation
Business Intelligence
KPI Development
Executive Reporting
Data Storytelling
Process Optimization
Projects
A business-driven case study where I performed in-depth EDA on the classic Northwind dataset. I uncovered key trends in sales, customer behavior, and product performance, and built a professional dashboard for storytelling using Power BI and SQL.
Tech Stack: SQL, Power BI, Pandas
A feature-driven analytics project where I identified critical drivers of loan defaults. I applied statistical analysis and visual storytelling to assist in better loan disbursement strategies.
Tech Stack: Python, Matplotlib, Pandas
""")
# Machine Learning Section (initially hidden)
with gr.Row(visible=False, elem_classes="section-container ml-section") as ml_section:
with gr.Column():
# Back button
back_from_ml = gr.Button("← Back to Home", elem_classes="back-button back-button-ml")
gr.HTML("""
Machine Learning
My machine learning expertise spans from traditional algorithms to deep learning systems that solve real business challenges.
I've built end-to-end ML pipelines that deliver measurable impact, combining the right models with appropriate feature engineering
techniques. I focus on creating solutions that are not only technically sound but also deployable, maintainable,
and integrated with business workflows. With a solid foundation in Python-based ML frameworks and cloud
deployment platforms, I develop models that generate actionable predictions and insights.
""")
gr.HTML("""
Skills
Frameworks & Libraries
TensorFlow/Keras
PyTorch
Scikit-Learn
XGBoost/LightGBM
ML Techniques
Supervised Learning
Unsupervised Learning
Deep Learning
Natural Language Processing
MLOps
ML Pipelines
Model Monitoring
Deployment Strategies
Version Control (DVC)
Cloud ML Services
AWS SageMaker
Google AI Platform
Azure ML
MLflow
Projects
Built a regression model to predict the chances of a student getting admitted to top universities based on academic profiles. The project includes feature importance analysis, model tuning, and a live demo deployed with Streamlit.
Tech Stack: Scikit-learn, Streamlit, NumPy
A custom-trained assistant that answers queries about my resume and portfolio using NLP and retrieval techniques. Built to simulate real-time interactions with hiring teams, this project showcases my ability to work with large language models and create practical AI applications.
Tech Stack: LangChain, OpenAI, Gradio
Developed an advanced multi-speaker audio processing system that performs speaker diarization, transcription, and summarization to extract meaningful insights from multi-speaker conversations.
Tech Stack: PyTorch, Hugging Face Transformers, Gradio
""")
# Computer Vision Section (initially hidden)
with gr.Row(visible=False, elem_classes="section-container cv-section") as cv_section:
with gr.Column():
# Back button
back_from_cv = gr.Button("← Back to Home", elem_classes="back-button back-button-cv")
gr.HTML("""
Computer Vision
I'm passionate about developing computer vision systems that can perceive and understand visual information in ways that benefit humans.
My experience spans from implementing state-of-the-art algorithms to deploying them in real-world scenarios. I've worked on projects
that enable machines to "see" and interpret their environment through image processing, object detection, and image classification.
I focus particularly on applications that improve accessibility and solve tangible problems, creating CV solutions
that operate efficiently even with hardware constraints.
""")
gr.HTML("""
Skills
CV Techniques
Object Detection
Image Segmentation
Feature Extraction
Image Classification
CV Libraries
OpenCV
PIL/Pillow
TorchVision
TF Computer Vision
Deep Learning for CV
CNNs
YOLO frameworks
Transfer Learning
Object Recognition
Applications
Accessibility Solutions
OCR/Document Analysis
Motion Tracking
Edge Deployment
Projects
Designed a system using object detection and OCR to help visually impaired individuals find products and navigate shopping aisles. Developed with real-time feedback on Raspberry Pi and OAK-D camera, this project demonstrates my commitment to creating technology that solves real accessibility challenges.
Tech Stack: YOLOv8, OpenCV, Raspberry Pi
An edge solution using Raspberry Pi to monitor and count vehicles at intersections, providing real-time traffic flow analytics. This project demonstrates efficient deployment of CV models on resource-constrained devices.
Tech Stack: YOLOv5, Raspberry Pi, OpenCV
""")
with gr.Row(visible=True, elem_classes="landing-section") as landing_section:
with gr.Column():
# Profile section with picture - using the actual image from data folder
try:
# Get the image as data URI
profile_img_uri = image_to_data_uri("data/My_photo.jpeg")
gr.HTML(f"""
AI/ML Engineer & Data Scientist
I'm a software engineer turned AI/ML practitioner with a strong foundation in Commerce and experience in ML, computer vision, and data analytics.
I blend business understanding with data-driven thinking to create real-world solutions. Currently open to roles in Data Science, Machine Learning Engineering, and Computer Vision.
Python
Machine Learning
TensorFlow
PyTorch
Computer Vision
Data Analytics
SQL
Power BI
My Specializations
""")
except Exception as e:
# Fallback if image cannot be loaded
gr.HTML("""
AI/ML Engineer & Data Scientist
I'm a software engineer turned AI/ML practitioner with a strong foundation in Commerce and experience in ML, computer vision, and data analytics.
I blend business understanding with data-driven thinking to create real-world solutions. Currently open to roles in Data Science, Machine Learning Engineering, and Computer Vision.
Python
Machine Learning
TensorFlow
PyTorch
Computer Vision
Data Analytics
SQL
Power BI
My Specializations
""")
# Cards Grid with proper structure
with gr.Row(elem_classes="cards-grid"):
with gr.Column():
# Data Analytics Card
gr.HTML('')
da_button = gr.Button("Data Analytics", elem_classes="card-button")
gr.HTML("""
""" + data_analytics_icon + """
Data Analytics
Data storytelling, insights extraction, interactive dashboards & business problem-solving
')
ml_button = gr.Button("Machine Learning", elem_classes="card-button")
gr.HTML("""
""" + machine_learning_icon + """
Machine Learning
Feature engineering, model training, deployment & automation pipelines
')
cv_button = gr.Button("Computer Vision", elem_classes="card-button")
gr.HTML("""
""" + computer_vision_icon + """
Computer Vision
Object detection, image recognition, edge AI & accessibility applications