These are the research projects I undertook during my M.Sc. and B.Sc. coursework, focusing on a variety of advanced topics in computer science and engineering.
Multi-hop Question Answering System for Institutional Data
Developed a system inspired by HotpotQA to handle multi-hop reasoning from diverse institutional data sources like Google Scholar and university websites. (Technologies: Python, Django, PyTorch, TensorFlow, Keras, NumPy, Pandas, Apache Spark (MLlib))
Performance Research and Optimization on Jython Compared to Python and Java
Conducted performance evaluation of Jython in comparison to Python and Java in distributed systems, focusing on execution time and efficiency. (Technologies: Amazon SageMaker, AWS Lambda, S3 Bucket)
Tuning the Carrier Sensing Range of IEEE 802.11 MAC
Optimized the carrier sensing range of IEEE 802.11 MAC to improve MAC performance in multi-hop ad hoc networks. (Technologies: GNS3, NS-2, NetworkX)
An Efficient Approach of Computing Double Cut and Join Distance for Genomes with Duplicate Genes
Developed a heuristic approach to accelerate the computation of the double cut and join distance for genomes containing duplicate genes. (Technologies: Google Cloud Platform (GCP), API Gateway, CloudWatch.
These are industry projects I developed across various sectors, focusing on AI, Machine Learning, and Deep Learning solutions to automate processes and enhance decision-making.
Aptitudo is an end-to-end AI-powered recruitment and assessment platform owned by Aptitudo Inc., Canada..
The platform supports the complete hiring lifecycle—from posting recruitment campaigns to creating assessments, generating AI-based questions (MCQ, short answer, long answer, coding, and video-based), inviting candidates, conducting exams, and publishing results with both automated and human evaluation.
Built on a scalable three-tier architecture comprising a Vue.js frontend, Java Spring Boot backend, and a Django-based AI server for modular, high-performance processing.
Integrates GPT-4o–driven AI assessment generation, SSO authentication with Google and LinkedIn, Stripe-based subscription and payment management, and secure external API integrations.
Includes advanced exam integrity and analytics features such as IP tracking, mouse movement monitoring, fraud detection, detailed reports, and audit logs for enterprise-grade recruitment workflows.
The BISDP Project is a World Bank–funded government initiative, and its Business Intelligence (BI) module is a government-grade analytics platform developed to support regulatory oversight and operational decision-making for multiple stakeholders.
Designed and implemented an enterprise BI system capable of secure data aggregation, optimized querying, and role-based dashboards, enabling actionable insights across UAT and Production environments for regulatory authorities.
Developed advanced predictive and analytical models including lapse prediction for customer retention, financial forecasting, fraud detection, and accident analysis using supervised learning, regression models, decision trees, and time-series forecasting techniques such as SARIMAX.
Delivered stakeholder-specific reports and executive dashboards supporting transparency, compliance, and policy decisions, while ensuring strong security controls, audit logging, authentication mechanisms, and reliable on-premises deployment.
Automated CRM Agent: Developed a scalable banking services chatbot using Dialogflow, Python Flask, and MySQL to serve concurrent customers.
Automated E-recruitment Agent: Created an online recruitment automation tool using Python Flask, Selenium, and MySQL to streamline HR processes.
Automated HR Agent: Implemented an NLP-based HR assistant with Python Flask and Dialogflow to handle employee queries via RESTful web services.
Automated Loan Origination and Approval: Developed a decision support system for loan approvals using Deep Learning models, Regression, Naive Bayes, and KNN algorithms.
ERP Data Predictive Analysis: Built predictive models for ERP data (sales, production) using Deep Learning techniques (Python Django, TensorFlow, Keras).
Online Credit Approval System: Developed a web service in Flask for online credit approval, integrating external APIs and databases (Python Flask, Heroku).