Shemonto Das
I am currently pursuing my Master's degree at the Department of Computer Science , Memorial University of Newfoundland , where I am focusing on developing Machine Learning techniques for time series analysis and classification as part of my thesis. My passion lies in the realms of data, analytics, and artificial intelligence, and I am always eager to delve deeper into these fields and expand my knowledge. I am seeking opportunities where I can apply my expertise and further enhance my skills in data science and AI. I want to obtain a position where my knowledge and experience can be utilized as well as enhanced. My resume is available here.
Experience
Data Science Intern
Graduate Teaching Assistant
Contractual Lecturer
Machine Learning Engineer
Team Lead, Communication and Control
President (Highest Executive Position)
Education
MSc. in Computer Science
BSc. in Computer Science and Engineering
Technical Skills
- Language and Frameworks: Python, Java, HTML, CSS, SQL, TensorFlow, Keras, Scikit-learn, Pandas, Numpy, Matplotlib, Tkinter.
- Software and Tools: VS Code, Google Colab, Anaconda Navigator, Git, Tableau, Arduino, Raspberry pi, Nvidia Jetson Nano, MS Office, MS Excel, Powerpoint.
Research
Research Interests
- Artificial Intelligence and Machine Learning.
- Data Science, Analytics and Visualization
- Deep Learning and Neural Networks.
- Bio medical image processing.
Publications
- Active Learning Strategies for Robotic Tactile Texture Recognition Tasks, Journal: Frontiers in Robotics and AI, section Robot Learning and Evolution
- Unbalanced Fault Classification Using Active Learning in Synthetic Fiber Manufacturing Process, IEEE Syscon 2024
Thesis
- An Active Learning Framework with Class Balancing Strategies for Time Series Classification (MSc. Thesis)
- An Active-Learning based Training-Schedule for Biomedical Image Segmentation on Deep Neural Networks (BSc. Thesis)
Projects
Project 3: Instrumar's Pipeline for Manufacturing Fault Classification
(Python, Scikit-learn, Pandas)
Project 4: Financial Fraud Analysis and Detection
(Python, Scikit-learn, Pandas, Imbalanced-learn)
Awards & Achievements
- Graduate Research Award 2023 for strong research contribution, Department of Computer Science, Memorial University.
- Awarded for Best Graduate Talk at the SEA 2023 conference by the Faculty of Science, Memorial University.
- Recipient of Vice Chancellor's Certificate For High Standard Of Discipline And Devotion During Residential Semester At , BRAC University [September - December 2017 / RS46].
- Secured 7th position at Singapore Autonomous Underwater Vehicle Challenge, 2018.
- Champion (Line following Robot), DUET Tech Fest 2018.
- Appreciation certificate awarded by Bangladesh Robotics Foundation
- Placed on Dean’s list for multiple semesters at BRACU and graduated with high distinction
- Champion, State level Inter School English Parliamentary Debate Competition
Recommendations
Vimal Simha, Data Scientist
Instrumar Limited
" I have known Shemonto Das for over a year, during his internship at Instrumar Limited, where I work as a Data Scientist. He played an active role in developing a project to automate fault detection in fibre manufacturing using Instrumar’s electromagnetic sensors. At Instrumar, we are running trials at a fibre manufacturing plant that incorporates the Active Learning strategies that Shemonto developed. Shemonto’s project was to use Active Learning strategies to identify and classify faults in the fibre manufacturing process and products using time-series data collected by Instrumar’s sensors deployed at fibre manufacturing plants. He tapped on his knowledge of Active Learning strategies to classify textures explored by robotic arms and researched existing to help us jointly formulate a plan for using Active Learning strategies to analyze Instrumar’s data. Shemonto also collaborated with our team to build a pipeline to process time-series data, extract relevant data features for his models and identify appropriate performance indicators and metrics to evaluate the models. The results of his project have now been accepted for presentation at IEEE Syscon. Instrumar’s dataset suffers from severe “class imbalance” i.e. there are large differences between the frequencies of different types of faults. To address this, Shemonto consulted his professors and the literature, and worked out a “class balancement technique” which he implemented. When we needed additional computing resources, Shemonto’s leadership and go-getter spirit ensured he found a way to use computational resources from the Digital Research Alliance of Canada and use their HPC clusters, without which completing the project would have been challenging. Shemonto is an outstanding data scientist and data engineer who thrives on challenges in his day-to-day work. His inquisitiveness and can-do attitude inspires many diverse approaches to problem solving. Come rain or shine or snow, which we get copious amounts of in Newfoundland, we could always count on Shemonto’s presence at our weekly meetings, in which he demonstrated excellent communication skills and a spirit of teamwork. If you are looking for an excellent data scientist or data engineer for your team, who is also a problem solver and a great team player, Shemonto is your pick"
This recomendation is provided in linkedin. link
Joshua Swamidas, Manager of Technical Solutions
Instrumar Limited
" Shemonto’s Masters project was to develop an Active Learning pipeline for our time series sensor data. He worked with out internal data scientist to analyze our data, provide timely feedback on what we needed to do to improve our data cleaning so that the data going into the pipeline would give us the best results. He was able to explain what he was doing, answer our concerns, and also explore different techniques when we started implementing this with two of our customers. He gained experience find out that implementing things in the wild is never straightforward and is much more intricate and involved with live data. The results from our first customer were excellent and helped us to uncover new features of interest. During his program, Shemonto interned with us to support our research, learn our domain, support our data analytics and software teams to develop data features that provided us better intelligence with the data and generate appropriate models from the data. Shemonto participated and contributed positively in most of our design conversations. He was able to direct and guide us when we didn’t understand his work. Shemonto also helped us to develop a presentation to explain the complex process to non-technical people. I would recommend Shemonto’s data-science, analytics and problem solving abilities for future and potential employers."
This recomendation is provided in linkedin. link
Contact
- shemontod [at] mun [dot] ca