Selected Awards
Valedictorian
North South University
Chancellor's Gold Medal
North South University
Honorary Award
2023 ICPC Asia Dhaka Regional Site Contest
NSU Full-merit Scholarship
North South University
Academic Mentor (2021)
ACM NSU
Experience
AI Engineer
2024 - PresentOptimizely
• Deployed five AI agents and copilots—including OptiGPT, now the daily driver for 80% of employees—using custom LLMs, RAG pipelines, and GenAI models, cutting support time by 40%.
• Reduced API costs by 30% through prompt engineering and optimizing data pipeline for faster model training and efficient resource utilization.
• Conducted workshops on ML infrastructure best practices, CI/CD automation, feature engineering, compliance, and LLM tuning, improving AI system reliability and reducing ML deployment errors by 25%.
AI Quality Analyst
2024 - PresentInvisible Technologies
• Improved NLP performance by 18% by leading 40 AI trainers to implement a rigorous testing and model evaluation framework, reducing error rates and enhancing reliability for enterprise applications.
Senior AI Data Trainer
2024Invisible Technologies
• Increased Large Language Model (LLM) response accuracy by 15% by fine-tuning AI models based on user feedback.
Research Scientist
2023 - 2024Mahdy Research Lab, NSU
• Trained 300+ researchers in Artificial Intelligence, Deep Learning, Computer Vision, Quantum Computing, and Medical Image Processing research, increasing trainees’ research project success rates by 30%.
Engineering Manager
2022 - 2023The Odd Pro
• Drove $500K in revenue by leading a data-driven engineering team to develop a SaaS platform with intelligent automation, reducing manual processing time by 50%.
• Increased operational efficiency by 40% using agile methodologies and ML-based automation solutions, eliminating redundant workflows across the software development life cycle and improving user engagement.
• Integrated generative AI to architect and optimize technically advanced and visually compelling product designs.
Software Engineer, Machine Learning
2020 - 2024Fiverr
• Delivered 40+ end-to-end ML, cloud computing, and full-stack enterprise software solutions to clients from 20+ countries with 100% success rates showcasing problem-solving, troubleshooting, and strong communication skills.
Chief Technology Officer
2016 - 2020CDPF
• Led cross-functional teams to develop a nationwide data-driven training platform used by 10,000+ students, increasing personalized learning effectiveness by 35% through AI-driven recommendation systems.
Publications
Predicting life satisfaction using machine learning and explainable AI
Life satisfaction is a crucial facet of human well-being. Hence, research on life satisfaction is incumbent for understanding how individuals experience their lives and influencing interventions targeted at enhancing mental health and well-being. Life satisfaction has traditionally been measured using analog, complicated, and frequently error-prone methods. These methods raise questions concerning validation and propagation. However, this study demonstrates the potential for machine learning algorithms to predict life satisfaction with a high accuracy of 93.80% and a 73.00% macro F1-score. The dataset comes from a government survey of 19000 people aged 16-64 years in Denmark. Using feature learning techniques, 27 significant questions for assessing contentment were extracted, making the study highly reproducible, simple, and easily interpretable.
Shadow: A Novel Loss Function for Efficient Training in Siamese Networks
In this paper, we present a novel loss function called Shadow Loss that compresses the dimensions of an embedding space during loss calculation without loss of performance. The distance between the projections of the embeddings is learned from inputs on a compact projection space where distances directly correspond to a measure of class similarity. Projecting on a lower-dimension projection space, our loss function converges faster, and the resulting classified image clusters have higher inter-class and smaller intra-class distances. Shadow Loss not only reduces embedding dimensions favoring memory constraint devices but also consistently performs better than the state-of-the-art Triplet Margin Loss by an accuracy of 5%-10% across diverse datasets. The proposed loss function is also model agnostic, upholding its performance across several tested models. Its effectiveness and robustness across balanced, imbalanced, medical, and non-medical image datasets suggests that it is not specific to a particular model or dataset but demonstrates superior performance consistently while using less memory and computation.
Quantum Energy Teleportation across Multi-Qubit Systems using W-State Entanglement
Quantum-energy teleportation (QET) has so far only been realised on a two-qubit platform. Real-world communication, however, typically involves multiple parties. Here we design and experimentally demonstrate the first multi-qubit QET protocol using a robust W-state multipartite entanglement. Three-, four- and five-qubit circuits were executed both on noiseless simulators and on IBM superconducting hardware. In every case a single sender injects an energy E0 that is then deterministically and decrementally harvested by several remote receivers, confirming that energy introduced at one node can be redistributed among many entangled subsystems at light-speed-limited classical latency. Our results open a practical route toward energy-aware quantum networks.