MANASWIMONDOL.
Doctoral candidate and Academic Associate at the University of Zurich. MSc in Data Science and Quantitative Finance from the University Zurich.
"When you want something, all the universe conspires in helping you to achieve it."
// The Alchemist - Paulo Coelho
About Me
私についてI am Manaswi Mondol, a Doctoral Candidate and Academic Associate at the University of Zurich, working under Prof. Dr. Charles Driver at the Department of Psychology on time series analysis, continuous time dynamical systems, and deep learning.
I hold a Master of Science in Data Science from the University of Zurich and ETH Zurich, with a minor in Banking and Finance. My academic journey began at the University of Twente (Netherlands), where I earned a Bachelor of Science in Business & IT. This foundation allowed me to specialize early in the intersection of technology and analysis, culminating in a thesis focused on the application of Machine Learning for earthquake prediction.
Like Joker in Persona 5, I believe in looking beyond the surface to find hidden truths. Whether it is cryptocurrency markets, healthcare analytics, or earthquake patterns, I apply deep learning and explainable AI to decode the complex signals in data.
"Fall seven times, stand up eight."
// Japanese Proverb
Languages
Persona 5
The rebel aesthetic, the pursuit of truth behind masks, and the power of transformation. Persona shaped how I see narrative and design.
The Alchemist
Santiago's journey taught me that the path matters as much as the destination. Every dataset is a Personal Legend waiting to unfold.
Japan
The precision of Japanese culture, the balance between tradition and technology. Wabi-sabi applied to the art of data science.
Data Science
MSc from UZH & ETH Zurich, now pushing boundaries as an Academic Associate. Deep learning, time series, and explainable AI.
Expertise
技術Languages & Frameworks
Machine Learning & AI
Data Engineering
Research & Methods
Tech Stack
Research
研究Time Series & Dynamical Systems
Modelling continuous time dynamical systems using deep learning. Time series analysis and forecasting under Prof. Dr. Charles Driver at UZH.
Explainable AI
Incorporating XAI techniques like saliency maps and integrated gradients to provide post-hoc explainability for deep reinforcement learning models.
Deep Reinforcement Learning
Implementing DQN, A2C, PPO, and DDPG algorithms for automated cryptocurrency trading with fractional trading and blockchain metrics integration.
Selected Publications & Theses
A Deep Reinforcement Learning Approach with Explainability for Cryptocurrency Trading
Master's Thesis - University of Zurich
How Many Stars Equal a Happy Emoji? A Comparative Analysis on Evaluation Bias Across Different Online Rating Systems
Master's Project - University of Zurich
Application of Healthcare Analytics and Machine Learning in Diabetes Prediction
Conference Paper - Information Systems Seminar, University of Zurich
Analysis and Prediction of Earthquakes Using Different Machine Learning Techniques
Bachelor's Thesis - University of Twente (Grade: 9.1/10)
Projects
作品Implemented DQN, A2C, PPO, and DDPG algorithms for automated BTC/ETH trading. DDPG achieved 107.95% annual return, significantly outperforming existing models. Integrated blockchain metrics and XAI techniques (saliency maps, integrated gradients) for model explainability.
Examined evaluation biases across different online rating systems (stars, emojis, sliders) in the United States. Discovered the 'ceiling effect' and 'primacy effect' in visual rating systems. Analyzed impact of demographics, education, and IT skills on ratings.
Trained Random Forest, Linear/Polynomial Regression, and LSTM models for predicting earthquake magnitude and depth. Polynomial regression showed best overall results for magnitude; Random Forest excelled at depth prediction. Bachelor thesis scored 9.1/10.
Other Projects
Sentiment analysis and topic modelling using BERT and clustering on 10,000 scientific articles to find relevant papers for research. Includes preprocessing, tokenization, and uniqueness scoring.
Tested Random Forest, SVM, and Logistic Regression on PIMA Indian diabetes dataset. Random Forest achieved best overall performance. Conference paper for Information Systems Seminar at UZH.
Journey
道Academic Associate - Research & Teaching
University of Zurich
Research under Prof. Dr. Charles Driver. Time series analysis, continuous time dynamical systems with Deep Learning, model optimization, explainable AI. Teaching seminars for psychology students and supporting statistical consulting.
Research Assistant
University of Zurich
Data analysis, web development, ML, topic modelling, sentiment analysis, machine translation, entity extraction. Support in teaching and independent research.
MSc Data Science
University of Zurich & ETH Zurich
Minor in Banking and Finance. Thesis on Deep Reinforcement Learning with Explainability for cryptocurrency trading. Courses in ML, AI, blockchain economics, quantitative finance.
Research Project - Earthquake Prediction
University of Twente
Independent research under Dr. Faizan Ahmed. Analysis and prediction of earthquakes using Random Forest, Polynomial Regression, and LSTM models.
BSc Business & Information Technology
University of Twente, Netherlands
Major in Business and IT, Minor in Applied Mathematics and Web Science. Bachelor thesis scored 9.1/10 on earthquake prediction using ML.
Quantitative Trading
Self-employed
Developed proprietary prediction models for market volatility with 75% accuracy over 2.5 years. Created trading strategies maximizing upside while minimizing risk.
Contact
連絡Let's turn data
into impact.
I am always open to research collaborations, data science projects, and opportunities that push the boundaries of knowledge. Whether you are working on time series, DRL, healthcare analytics, or NLP, I would love to connect.
"There is only one thing that makes a dream impossible to achieve: the fear of failure."
// The Alchemist - Paulo Coelho