Jyoti Rani

· Ph.D. candidate with expertise in advanced machine learning, deep learning, and generative Artificial Intelligence (AI) algorithms.

To seek and maintain full-time position that offers professional challenges utilizing interpersonal skills, excellent time management and problem-solving skills.

DOCTORAL THESIS

Title: Data-driven Process Anomaly Detection Using Machine Learning and Deep Learning Techniques

Supervisor: Dr. Hariprasad Kodamana, Associate Professor, Department of Chemical Engineering and School of Artificial Intelligence, IIT Delhi, India.

Description:

· Developed a novel data-driven fault detection model integrating the Hidden Markov model(HMM) with Probabilistic Principal Component Analysis (PPCA) and Dynamic Principal Component Analysis (DPCA).

· Introduced a pioneering reconstruction-based fault detection algorithm, Generative Adversarial Autoencoders (GAAE), tailored for time series data.

· Conceptualized and executed a cutting-edge fault detection and isolation model using a probabilistic wavelet neural operator auto-encoder (PWNOAE) with application to dynamic processes.

· Innovated a novel distribution learning-based fault detection algorithm, Generative Adversarial Wavelet Neural Operator (GAWNO).

· Developed a specialized Probabilistic Fourier Neural Operator (PFNO) designed for advanced fault detection applications.

MASTER THESIS

Title: Chemical Looping with Oxygen Uncoupling (CLOU) of high ash and low ash coal using Co3O4 under N2 and CO2 atmosphere.

Supervisor: Dr. Prabu Vairakannu, Associate Professor, Department of Chemical Engineering, IIT Guwahati, India.

Description: Chemical Looping with Oxygen Uncoupling (CLOU) using Co3O4 was performed for the inherent capture of CO2 gas for both high-ash and low-ash coal. The net thermal efficiency of the power systems is estimated for the CLOU system with Co3O4 as an oxygen carrier. (Experiment with simulation using ASPEN plus).

• Title: Data Scientist
• Date of Birth: 09/26/94

· Generative Adversarial Wavelet Neural Operator (GAWNO)

Engineered GAWNO using advanced AI algorithms, achieving a 14% accuracy boost in fault detection over traditional GANs.

· Developed PWNOAE - Multivariate Process Data Distribution Learning Model

Developed PWNOAE, an innovative multivariate process data distribution learning model, outperforming Autoencoders, LSTM, and RNN-based algorithms with an impressive 8% performance enhancement.

· Fault Detection of Pressurized Heavy Water Nuclear Reactors (PHWR)

Collaborated with NPCIL team, achieving a 16% improvement in fault detection efficacy by integrating HMM, PPCA, and DPCA in a data-driven approach over traditional models like K-Nearest neighbors (KNN), Principal component analysis (PCA) and Support Vector Machine (SVM)

· Generative Adversarial Auto-Encoders (GAAE) for Time Series Fault Detection

Developed GAAE to address complex time series process data faults, showcasing a 13% efficacy improvement through expert evaluation and refinement of AI models like Autoencoders, LSTM, RNN, GANs, and PCA.

· Hands-on Training Program on ASPEN plus 2018 in IIT Guwahati.

· Hands-on Training Program on MATLAB 2018 in IIT Guwahati.

· Hands-on Training Program on Python at IIT Delhi.