Sam Narayana Subudhi

Looking for Data Science Roles | ML Enthusiast

Summary

Final year Data Science Postgraduate pursuing Masters at Indian Institute of Technology, Palakkad . Interested in developing Python and Deep Learning projects.

Web Development

Course Solutions

Code Snippets

Machine Learning


My Skills

Few of the skills that I consider myself acquainted with, to serve the given opportunity are showcased here.

C

70%

Python

80%

MySQL

80%

GitHub/Git

60%

Machine Learning

80%


Work Experience

MAQ Software ,Hyderabad May 2023 - July 2023

Associate Software Engineer

  • Gained valuable experience working with a dynamic and innovative team.
  • Dived into various technologies gaining hands-on experience with Power BI, Azure, and other Microsoft tools.
  • Actively contributed to projects, working for a leading MNC client, enhancing technical, problem-solving, and teamwork abilities.
  • Worked on data visualization initiatives using Power BI to create dynamic dashboards, facilitating data-driven decision-making.
  • Specialized in data validation and ensured data integrity by extracting and cleansing data from source and mart tables using SQL queries. This critical role in the data pipeline underscored my commitment to accuracy and quality in data analysis, supporting informed decision-making processes
  • Utilized Azure cloud services, including Azure Notebooks, to streamline data processing and analysis workflows.

Tata Consultancy Services, Bhubaneswar September 2020 - June 2022

Systems Engineer

  • Developed and maintained efficient PL/SQL procedures, functions, and packages, boosting data processing for a Leading MNC client .
  • Worked within an agile team and helped prioritize and scope feature requests to ensure that the biggest impact features were worked on first.
  • Designed and implemented efficient workflows, improving data communication and integration for analytics purposes .
  • Worked on data visualization initiatives using Power BI to create dynamic dashboards, facilitating data-driven decision-making.
  • Partnered with QA to ensure data integrity, contributing to the reliability of data analysis outcomes, identifying and addressing software defects and ensuring high-quality deliverables.
  • Crafted advanced SQL queries to facilitate effective data analysis, supporting strategic decision-making with insightful data retrieval.

Projects

Low Quality Face Recognition in Automatic Attendance Systems (Ongoing)

  • Developed a Classroom Automatic Attendance System (CAAS) using a camera mounted in the classroom by capturing photos and videos of live classrooms and process images for recognizing the students present in the class using machine learning techniques
  • The system will detect faces and extract them from classroom images, and then compare them with a database of faces collected during course enrolment to register attendance automatically.
  • An App is being created, allowing users to upload images or use a webcam for Automatic Attendance

Activity Recognition task on UCF 101 Dataset

  • Utilizied various machine learning models to accurately classify video clips from the UCF 101 dataset. The goal was to achieve high accuracy in classifying the video clips into different action categories .
  • Achieved high accuracy,developed and implemented multiple models, worked collaboratively, conducted evaluation of models, and identified potential areas for further improvements.

Dashboard for Data Analysis

  • The project included preprocessing data using NumPy, Pandas, embed SQL to python scripts using SQLalchemy and construct a dashboard using streamlit to access data, analyze, visualize and infer .
  • Implemented Streamlit-based website allowing users to edit the data and see the change in visualizations.

Real-Time Human Counting

  • Used YOLOv5 models for real-time human counting, emphasizing model development and comparison across diverse datasets to ensure robustness and accuracy .
  • A Flask-based web application was created, allowing users to upload images/videos or use a webcam for live counting.

COVID-19 Death Prediction

  • Developed a Predictive Model for COVID-19 Fatalities : - Utilized several ML Models to analyze global week-by-week pandemic data, accurately forecasting weekly death tolls by region. This involved sophisticated data preprocessing and modeling techniques to handle complex datasets effectively.
  • The predictive insights generated can be used in assisting health authorities in strategic planning and resource allocation, demonstrating the model’s utility in guiding critical public health decisions during the pandemic.

Remote Homology Detection Using Physiochemical Properties

  • This work presents a novel method to use physicochemical parameters and frequency density for feature vector representation. This method was assessed using popular machine learning paradigm to classify protein sequences for predefined homology.

Contact Me