I am a performance-oriented Computer Science graduate specialized in AI and ML from Vellore Institute of Technology, Chennai, India, currently pursuing an MSc in Business Analytics from Trinity College Dublin.
I am a performance-oriented Computer Science graduate specialized in AI and ML from Vellore Institute of Technology, Chennai, India, currently pursuing an MSc in Business Analytics from Trinity College Dublin. I have academic research experience at IIT BHU, NUS, Duke University, and NIE (NTU Singapore), with focus on machine learning, data analytics, and business intelligence. I have contributed to financial risk prediction, fraud detection, and health diagnostics, and I enjoy synthesizing technical skills with business insights to drive data-driven strategies.
I thrive at the intersection of research and application, bridging complex analytics with practical business outcomes. I enjoy building end-to-end ML pipelines, collaborating across disciplines, and applying data science to finance, healthcare, and technology domains to create measurable impact.
Constructed deep learning-based models for corporate financial distress prediction, utilizing advanced time series techniques to support early risk identification for investors and creditors.
Gained high performances with advanced deep learning architectures — Transformers and Attention-based LSTM models attained over 95% accuracy.
Designed a time-series forecasting platform analyzing exchange rate volatility using ARIMA and GARCH models, supported by statistical testing (ADF, KPSS, MAE, RMSE) to strengthen risk assessment and forecasting accuracy.
Designed an analytics platform using Python and Tableau to analyze 77K+ YouTube videos (2020–2024), uncovering insights into engagement, audience behaviour, and ad performance.
Built and deployed an ML-based web app that predicts relevant YouTube channels from video titles and tags, enhancing content discovery and marketing strategies.
Architected an automated malaria detection pipeline on RBC smear images using EfficientNetB3, InceptionV3, and VGG16; optimized diagnostic precision through transfer learning and weighted aggregating and deployed it on the web using Flask Framework.
Attained model performances with EfficientNet B3 (98.76% accuracy), VGG16 (94.45% accuracy), and InceptionV3 (85.34% accuracy).
Devised and validated ML/DL frameworks under NUS supervision to detect Ethereum-based fraud, conducted blockchain data exploration and engineered key features to enhance predictive accuracy.
Cooperated on a research project under the supervision of NUS faculty, focusing on blockchain, Cryptocurrency, Decentralized Finance, Exploratory Data Analysis, Data Science and Deep Learning.
Extended IIT BHU research on EEG-based epileptic seizure detection using STFT, CWT, Attention-LSTM, and Transformer models, improving classification robustness and evaluation metrics (ROC-AUC, confusion matrix).
Collaborated with PhD students to develop deep learning models (Transformer Networks, Attention-based LSTM) and applied time-frequency spectrogram techniques (CWT, STFT) for improved classification of epileptic seizures from pre-processed TUH EEG dataset, evaluated using graphical representations, confusion matrices, and performance metrics.
Post a job and invite Yashvardhan, or contact them directly. No hidden commission.
