Ramsundar Mahato
Hi there, I am Ramsundar, I'm a Data Scientist at Thoucentric, where I ship my boat scale.
I am a growing Data Science practitioner changing physically, searching for ways to attain complete mastery of my discipline. With having 4.5+ years of experience in executing data-driven solutions to increase efficiency & utility of internal data processing and also having Experienced at creating different ML, DL models to deliver insights and implement action-oriented solutions to complex business problems.
I also working on few new practices and can be viewed here Github.
Experience
Data Scientist I
Collaborates with product teams to build, deploy ML systems for strategic planning and cornerstone for other AI,ML products. Design and building a classification system to monitoring pattern change and anomaly detection in a systematic way. Work closely with production team to build up a streamline data science platform. Improving efficiency on whole ML cycle including data pipeline, model building, serving and monitoring.
Technology: Python, Statistics, Machine Learning, Advanced Excel, Keras, MS PowerPoint.
- Collected, Studied and Interpreted large datasets; conducted a report and performed accurate and successful data management.
- Collaborated with all the stakeholders to minimize the warehouse and plant loss for the largest volume brewery firm..
- Performed Exploratory Data Analysis to investigate the client’s data to spot anomalies, and patterns, as well as to test the hypothesis made, based on the statistical metrics and visualizations.
- Created a customized Dashboard for a data source for reporting sales, inventory, and marketing..
Associate Data Scientist
Demand Forecasting of Empty Containers and Optimizing Route to Demand Origin
Technology: Linear programming, Python, Data Pipelining, DOMINO, Machine Learning, XGB, Random Forest, MS PowerPoint.
- Utilized, and deployed XGBoost Model for Forecasting 60,90,120 days for all the DFUs and predicting the demand from historical data
- Applied Linear Programing and Ant Colony Optimization Theory for identifying a feasible solutions for all routes (P-to-P)
- Built Optimization Model to maximize the utilization of containers and reduce the cost of transportation
- Code refactoring and packaging (PyPi Modules-whl file) for the production pipeline on Bitbucket repo.
- Integrated API connections for ADLS (Data Lake), UI and Dremio database.
- Created Databricks Jobs with whl file and connected to Airflow DAG.
- Dockers container setup for Airflow-DAG and implemented scheduled execution of Task-Flow for Databricks Jobs.
- Active participation in Sprint meetings and valuable contributor
Data Scientist
Developed a machine learning model to predict the likelihood of a store selling illicit goods.
Technology: ML Algorithms, Modelling, EDA, BitBucket, Time Series Analysis, OOPS, Excel, AWS S3, Git, Confluence,
PowerPoint.
- Developed the end-to-end pipeline for the illicit trade prediction project, encompassing data collection, preprocessing, feature engineering, model training, and deployment
- Performed multivariate analysis to gain insights into the data and identify correlations between various factors such as store level, region, barangay, and illicit cigarette presence
- Made informed assumptions about sales and survey data to address data sparsity, in accordance with business requirements
- Utilized the LightGBM model, a gradient boosting framework, to build a predictive model capable of forecasting the likelihood of a store containing illicit cigarettes based on historical data patterns
- Engineered complex features to establish correlations between geographical factors and sales, enabling a more accurate prediction of illicit trade presence. Also created features based out of high/low selling stores
- Documented the project, including methodologies, data preprocessing steps, feature engineering techniques, and model performance evaluation, to provide a comprehensive understanding of the illicit trade prediction pipeline
- Re-factored and packaged the code using OOPs concepts, creating a PyPi Modules-whl file for the production pipeline on Bitbucket
Software Developer Intern
Developed a Payroll System for the company during the internship period.
Technology: Java, Servlet, Session, JSP, SQLite, Spring.
- Used different spring components like Spring Core, Spring JDBC to manage business layer.
- Proficiency in Java, J2EE (JSP, SERVLETS and REST Web services) and frameworks like Spring MVC).
- Versatile team player and individual contributor with strong analytical and problem-solving skills
- Creating database schema that represent and support business processes.
- Mentor other team members and review their deliverables.
Community
Youtube
Experienced mentor specializing in teaching software development, machine learning, artificial intelligence (AI), and leveraging Natural Language Processing (NLP) and LLM terminologies. Proficient in tutoring software development at a production level, encompassing the entire software lifecycle. Skilled in technologies like Java, Swing for GUI, jTattoo for Look and Feel, SQLite for Backend Database, alongside expertise in Machine Learning (ML) concepts. Capable of explaining every line of code in machine learning projects.
Also contributed lot's of ML Based projects with explained each and every line of codes.
Actively engages in answering questions within discussion forums, boasting over ~9K subscribers.
Education
IIIT Bangalore & Liverpool John Moores University
- Published Paper based on....
Uttrakhand Technical University
- Got 1st in ProTyping organized by UTKRISHT 2017.
- Got Letter of Recommendation from Director, UTU for Developing Website.
- Participated in Smart India Hackathon 2017 & 2018.
- Got selected for making Human and Object Detection for Final year Project.
Kendriya Vidyalaya
- State Level Cricket Player
- District Level Badminton Player
- Athletes
Volunteer Experience
Website Developer
Honored for developing website for our college from Mrs. Silky Jain (Executive Director ) and Anuj Singh (HOD CSE Dept.) on the occasion of UTKRISHT 2018 the Technical Fest.
- Designed Wire frames along with the UI Design Team.
- Developed Front End Pages Using HTML, CSS and JavaScript, Web design and development
- Developed various User Forms, GUI Components using JSP, HTML
I am also responsible for writing server-side web application logic. I usually develop back-end components, connect the application with the other (often third-party) web services, and support the front-end developers by integrating their work with the application.
Student Volunteer
Each one Teach One Initiative and helping more fellow Indians in becoming the part of digital Economy.
Skills
Languages: Python, Java, Nodejs
Cloud Computing: Azure, AWS, Azure ML, Azure Databrick, Sage Maker, Cognitive AI, Azure Functions
Key Skills: Machine Learning, Deep Learning, NLP, LLM Prompt Design, Predictive & Statistical Modelling, GenAI, Flask, Web Development, Timeseries Modelling, API Building, Logic Apps, Power Apps, Project & Delivery Management, Strategic Mindset
Database: SQL & Postgres
Tools: OpenAI, Power BI, Excel, DOMINO, HON Forge, Power Platform, Jupyter NB, VS Code, Confluence, MLFlow, BitBucket
My Passions
- Capturing life's moments through photography 📸
- Whipping up delectable treats in the kitchen 🍪
- Engaging in friendly board game battles 🎲
- Gathering around crackling campfires under the stars 🔥
- Experimenting with new culinary creations 🍴
- Embracing the tranquility of nature through hiking trails 🌲
- Embarking on spontaneous midnight drives 🚗
- Keeping fit and energized through running 🏃♂️
- Unabashedly enjoying the tunes of Taylor Swift and Ariana Grande 🎶
- Expressing my creativity through writing and storytelling ✍️
created by Me