• Passionate data scientist and ML Engineer with over 6+ years of experience, deeply interested in the transformative power of data.
• Proficient in mathematics, machine learning, deep learning, real-time analytics, and algorithm optimization.
• Enthusiastic about understanding the underlying principles behind algorithms and their practical applications.
• Thrives in collaborative environments and enjoys exchanging knowledge with motivated peers.
• Fascinated by technology, emerging trends, and the intersection of art and design.
• Dedicated to continuous learning and exploring philosophical concepts to broaden perspectives

Programming Languages: Python, Javascript, SQL, PySpark, R
Machine Learning and Deep Learning: Tensorflow, PyTorch, SciPy
Model Deployment: Torchserve, Tensorflow Server
Image Processing: OpenCV, PIL

NLP: LLM, NLTK, Spacy
Web Automation and Scrapping: Selenium
Data Visualization Tools: Matplotlib, Seaborne
Web Technologies and Frameworks: Django, FastAPI
Databases: MySQL, PostgreSQL (SQLAlchemy), CosmosDB (MongoDB)
Mobile Technologies: Android (Basics)
Network Traffic Monitoring Tool: Nmap
Blockchain Technologies: Hyperledger Fabric (GoLang)
Deployment: GitHub, BitBucket, Heroku, Docker, Azure (App Service)
Others: LangChain (ChatGPT), TwitchIO, Raspberry Pi, Streamlit, Postman, CI/CD
Cloud Platforms: AWS EC2,S3, Lambda, SageMaker, Azure Data Bricks, Azure Data Factory
Data Visualization: Tableau, Power BI

• Coursera MOOC: Neural networks and Deeplearning - Deeplearning.ai

• Probability and Statistical Inference
• Data Analytics Fundamentals
• Technical Writing for Data Scientists
• Data Analytics and Visualization Techniques
• Data Mining and Knowledge Discovery
• Machine Learning Foundations
• Algorithm Optimization for Big Data
• Finite Automata and Formal Languages
• High-Performance Computing for Data Science
• Embedded Systems Design and IoT Applications

YORO (Financial and Landing Domain):
Tools and Technologies: Python, NLP, Django, Rest API, Seaborn, PyFPDF, Docker, Azure, GPT3 / 4
Project Description: YORO is a comprehensive financial and lending platform designed to streamline financial services and lending processes. The platform leverages advanced technologies to provide a seamless user experience, from loan application and approval to financial analytics and reporting.
Roles and Responsibilities:

• Automated loan application processing using Natural Language Processing (NLP) to extract and validate information from submitted documents.
• Used advanced data visualization techniques provided clear insights into financial trends, lending performance, and risk assessment.
• Developed secured and efficient RESTful APIs to facilitate seamless integration with Odoo.
• Leveraged the power of GPT-3/4 for predictive analytics, customer support, personalized financial
  recommendations, and natural language understanding (NLU) capabilities.

Churn Prediction:
Tools and Technologies: Python, Pandas, Numpy, Seaborn, Scikit-learn, TensorFlow, Django, Docker, Git, PyCharm
Project Description: Developed churn prediction model to identify customers at risk of leaving the service, enabled proactive retention strategies. It allowed the business to implement targeted retention strategies, thereby reducing churn rates and improving customer retention.
Roles and Responsibilities:

• Identified customers who are likely to leave (churn) the system, specifically focusing on credit card and loan customers.
• Developed a robust machine learning model capable of accurately predicting customer churn.
• Model was capable of giving insights into the key factors driving churn, aiding in the development of targeted retention strategies.
• This solution can be integrated into the any organization’s customer relationship management (CRM) system to provide real-time churn predictions.

OCR for Automated Invoice Processing
Tools and Technologies: Python, TrOCR, Scikit-Learn, SQL for Database Management, GPT3 / 4
Project Description: Developed an automated Optical Character Recognition (OCR) and machine learning to extract the details from the invoice and process invoice data efficiently, reducing manual intervention and improving accuracy. This project was to built a mechanism that can calculate the Rewarding Points basis the counter sale from retail outlets.
Roles and Responsibilities:

• Automated the extraction of key information from scanned invoices using OCR and machine learning techniques.
• Template-Independent design implemented with GEN AI algorithms to recognize and extract data regardless of invoice layout.
• Incorporated rule-based systems for validating extracted data against known formats.
• Provided seamless integration with existing accounting software for automated data entry.
• Generated synthetic documents based on the description and templates of various types using GPT3 and GPT4.

CANVA8
Tools and technology: Pytorch, LLM models like gpt2, BART, BART50, LAMMA for code generation,
Description: Project aims to automate the conversion of Figma design files into HTML/CSS code using advanced AI models and machine learning frameworks. The solution leverages PyTorch and large language models (LLMs) like GPT-2, BART, BART50, and LAMMA for accurate and efficient code generation.
Roles and Responsibilities:

• Extraction of design elements and properties from Figma files. Structured and organized the extracted data for further processing.
• Utilized PyTorch to build and train neural networks for understanding design-to-code translation.
• Fine-tuned pre-trained models such as GPT-2, BART, and LAMMA to enhance their capability in generating HTML/CSS from design inputs.
• Developed algorithms to interpret design elements and generate corresponding HTML/CSS code. Ensured the generated code is clean, maintainable, and follows best practices.

Educational Chat Bot
Tools and technology: Python, PyTorch, MongoDB, LLM, BART, BART50, LAMMA for content generation
Project Description: Project involved developing an advanced conversational AI system designed to assist learners by providing real-time, contextually relevant information and personalized support. The primary goal was to create an intelligent chatbot capable of understanding and generating educational content to enhance the learning experience.

Roles and Responsibilities:

• Developed an interactive educational software where students can engage with various applications, including a Chat Bot, with regional language.
• Incorporated Natural Language Processing (NLP) frameworks for the Chat Bot, enabling text and voice interaction.
• Utilized language translation APIs to support regional languages and enhance accessibility.
• Integrated Retrieval-Augmented Generation (RAG) model for generating truthful answers in both text and audio formats using GENAI model.
• Fine-tuned the RAG model on educational datasets to ensure accurate and contextually relevant responses.
• Implemented Text-to-Speech (TTS) libraries for converting textual responses into natural-sounding audio.
• Enhanced audio processing to maintain quality and improve user experience.
• Software was made equipped to generate Questions & Answers for different difficulty level.

Document Understanding
Tools and technology: Python, AWS Texract, TrOCR, LLM3, Donut Model
Project Description: It delivered advanced solutions tailored to address a spectrum of challenges associated with document management. AWS Textract for form data extraction for document understanding, TrOCR for ocr for printed and handwritten receipts and documents. LayoutLM3 for text entity classification. Donut Model direct text extraction
for document classification, text extraction, and document question answering
Roles and Responsibilities:

• Developed and fine-tuned models to categorize documents based on their content and purpose.
• Ensured seamless integration of classification models with the overall document management system.
• Used technologies like TrOCR to extract text from printed and handwritten documents.
• Prepared documents for OCR by preprocessing (e.g., image enhancement, noise reduction) to improve text extraction accuracy and applied algorithms to correct OCR errors and refined extracted text for further use.
• Analyzed the OCR-extracted text to group related documents into clusters or batches and Created algorithms to automate the grouping process based on document content, metadata, or other business criteria.

Dental Application
Tools and technology: Python, Django, Flask, Rest API, Mask R-CNN, Unet
Project Description: Developed and deployed cutting-edge machine learning solution. Collaborated closely with cross-functional teams to enhance our desktop applications with custom algorithms and integrated machine-learning models into our systems. My role was to involve developing Flask APIs for seamless integration, ensuring smooth functionality and performance of the application. This offered an exciting opportunity to work at the forefront of technology, solving complex problems and driving innovation in machine learning.
Roles and Responsibilities:

• Created and maintained APIs using Flask to enable seamless integration of machine learning models with the desktop application.
• Integrated custom algorithms and machine learning models into the application, enhancing its functionality and performance.
• Worked closely with other teams, such as data scientists, software engineers, and product managers, to ensure
  successful implementation and integration of machine learning solutions.
• Ensured the smooth operation of the application by optimizing the performance of both the machine learning models and the APIs.
• Addressed complex technical challenges and contributed to driving innovation within the application.