Nabil Abdullah

Designed QEC encoder and decoder code for 1 bit flip and 2 bit flip errors as well as 1 phase flip and 2 phase flip using quantum gates. Implemented Quantum Approximate Optimization Algorithm (QAOA) using quantum circuit for the combinatorial optimization problem of MaxCut.

The project focused on accelerating the Bitonic Sort algorithm using CUDA programming to achieve significantly higher performance compared to its CPU implementation. To achieve substantial speedups, several low-level optimizations were applied, particularly related to memory management strategies and the selection of efficient data types.

Trained three separate reinforcement learning models using Proximal Policy Optimization (PPO) for Time Trial, Object Avoidance, and Head-to-Head race types in AWS DeepRacer. Designed and implemented custom reward functions to optimize vehicle behavior specific to each race type, focusing on speed, stability, obstacle avoidance, and opponent interaction.

Built a KV-MemNN model from scratch for question-answering over natural language documents. Developed custom training and testing pipelines with integrated attention mechanisms for interpreting questions.

Built a KV-MemNN model from scratch for question-answering over natural language documents. Developed custom training and testing pipelines with integrated attention mechanisms for interpreting questions.

The project involved performing hyperparameter tuning on smaller existing multimodal models for hate speech detection in memes, with the goal of improving their accuracy and comparing their performance against larger models such as Google Gemini and GPT-4. Appropriate prompt engineering was necessary for the large multimodal models to ensure that their outputs were consistent and well-structured.

Evaluated four randomized optimization algorithms—Hill Climbing, Simulated Annealing, Genetic Algorithm, and MIMIC—on three different problem domains. Highlighted the comparative advantages of each algorithm based on convergence speed and solution quality.

The project involved selecting two classification problems and implementing models including Decision Trees, Boosted Decision Trees, SVM, KNN, and Neural Networks. Hyperparameter tuning was performed to improve accuracy, and the best-performing models were evaluated using test data.

Designed a PID controller for a drone to perform vertical lift, position hold, and horizontal motion through roll control. Tuned proportional, integral, and derivative gains to balance stability and responsiveness and performance evaluated through trajectory tracking and hover accuracy.

Spearheaded the modernization of legacy ASPX applications by transitioning to Angular, boosting system scalability by 30% and reducing page load times by 20%. Engineered dynamic, high-performance web applications by implementing front-end enhancements in Angular and C#, resulting in a 15% increase in user engagement and improving response times by 25%.

Deployed Python-based deep learning pipelines using TensorFlow to process and classify LiDAR-scanned terrains, reducing manual labeling time by 40% and enabling automated recognition of trees, buildings, and ancient ruins. Optimized data processing workflows, accelerating training times for large-scale LiDAR datasets by 25%.

Specialized in Machine Learning, focusing on deep learning, NLP, reinforcement learning, and GPU-accelerated computing. Built multimodal AI pipelines and deployed RL simulations in the cloud.

Completed a Bachelor of Technology in Instrumentation and Control Engineering. Selected as a campus recruit at Boeing, where I modernized legacy systems.

Completed a Bachelor of Technology in Instrumentation and Control Engineering. Gained foundational knowledge in engineering principles and systems.