Badges
Certifications
Work Experience
Data Scientist
WPDeveloper• June 2021 - Present
Applicant Tracking System (Resume Parsing & Resume Scoring) for Easy.Jobs Constructed a language model to accomplish the Named Entity Recognition which consisted of a multilayer neural network to predict whether a given word represented the categories : Name, Email, Phone Number, Address, Degree, College Name, Results, Company, Designation, Duration, Soft Skills, Technologies. The developed resume parser has been integrated with existing ATS systems, CMRs, software using NLP, Spacy, BERT, Regex, Tensorflow, Theano, Python, Flask, Named Entity Recognition (NER), OCR. Deployed Applicant Tracking System (ATS) to Easy.jobs saas product where I have used large-scale autoregressive language models (GPT3, GPT J, GPT Neo, Bert, XlNet etc) ; provided faster and better CV screening. Applied various transfer-learning techniques using pre-trained word-embedding like Glove, fastText, BERT, ELmo, Universal Sentence Encoder for text similarity tasks for resume scoring. Intelligent Recommendation System for Templately.com Engineered a Intelligent Recommendation System to provide most related templates to users. Deployed a recommendation engine to production to recommend templates at tamplately.com using Pandas, Matrix Factorization, Restricted Boltzmann Machines, Clustering and LSH, Nearest-neighbours, TF-IDF, Classifiers (e.g. ANN or Naive Bayes), Tensor Factorization, ML &; DL , increasing average download size by 14%. E-commerce Classifier Model for Shopify Researched, prototyped (from research paper), built features and optimised (hyper-parameter tuning) the state-of-the-art machine learning and deep learning techniques like SVM, Logistic Regression, Random Forest regression, LSTM, CNN etc., using scikit-learn, keras, tensorflow on CPU/GPU for E-commerce classifier model. Implemented CI/CD/CT pipelines for E-commerce classifier model. Automated the data science platform by building and managing data pipelines and extensions that bridge NoSQL-, APIs, Machine Learning Engine and UIs Collaborate with Google Cloud Platform (GCP) and Heroku to enhance the E-commerce classifier model. Shopify String Matching Model Applied Data Structure Algorithm (DSA) using Pandas for matching large amounts of string to reduce the execution time (reduced from 15-18 seconds to 0.4 second).
Education
Universiti Teknologi Petronas
Computer Information Systems, MS• August 2018 - March 2021
Journal Papers Corresponding/ 1st Author 1. “Optimization of the hydropower energy generation using Meta-Heuristic approaches: A review,” Energy Rep., vol. 6, pp. 2230–2248, 2020. Published in Energy Reports Q2 (ISI Q2, Scopus Q1, IF = 3.56) DOI: 10.1016/j.egyr.2020.08.009 2. “Artificial Intelligence approach to total organic carbon content prediction in shale gas reservoir using well logs: A review” (Published in International Journal of Innovative Computing, Information and Control (IJICIC)) (Q2 ISI: IF 0.44/ Q2 Scopus). DOI: 10.24507/ijicic.17.02.539 3. “A Modified Niching Crow Search Approach to Well Placement”. Published in Energies MDPI (Q2 ISI IF = 2.7/ Q2 Scopus). DOI: 10.3390/en14040857 4. “Feature Selection-Based Artificial Intelligence Techniques for Estimating Total Organic Carbon from Well Logs,”. Published in Journal of Physics: Conference Series Q3 (SCOPUS Q3, IF = 0.23) DOI: 10.1088/1742-6596/1529/4/042084 5. “Evaluation of Tree-Based Ensemble Learning Algorithms to Estimate Total Organic Carbon from Wireline Logs”. (Accepted in International Journal of Innovative Computing, Information and Control (IJICIC)) (Q2 ISI: IF 0.44/ Q2 Scopus). Co-author 6. “Modelling and optimization of microhardness of electroless Ni-P-TiO2 composite coating based on machine learning approaches and RSM”. Published in Journal of Materials Research and Technology (ISI Q1: IF= 5.289) DOI: 10.1016/j.jmrt.2021.03.063 7. “Comparison of Machine Learning Classifiers for Accurate Prediction of Real-Time Stuck Pipe Incidents,” Energies, vol. 13, no. 14, p. 3683, 2020. Published in Energies MDPI Q2 (ISI IF = 2.7) DOI: 10.3390/en13143683 8. Missing well log data handling in complex lithology prediction: An NIS apriori algorithm approach. Published in International Journal of Innovative Computing, Information and Control (IJICIC) (Q2 ISI: IF = 0.44). DOI: 10.24507/ijicic.16.03.1077 Conference Papers 1. Estcon ICER2020 Evaluation of TOC Using Hybridization of Deep Neural Networks and Genetic Algorithms. Book Chapter 1. M. S. A. Rahaman and P. Vasant, “Artificial Intelligence Approach for Predicting TOC from Well Logs in Shale Reservoirs: A Review,” in Deep Learning Techniques and Optimization Strategies in Big Data Analytics, IGI Global, 2020, pp. 46–77. DOI: 10.4018/978-1-7998-1192-3.ch004.