Autonomous Systems & Deep Reinforcement Learning Researcher
I am a graduate researcher with a strong background in Mechatronics and Control Theory. My research focuses on developing intelligent, provably safe decision-making algorithms for autonomous vehicle navigation, multi-agent trajectory forecasting, and deep generative computer vision in dynamic urban environments.
Formal academic training in Mechatronics Engineering, Control Theory, and Machine Learning.
Developing robust algorithms for perception, multi-agent forecasting, and optimal decision-making in dynamic environments.
Designing hybrid Reinforcement Learning (RL) and Imitation Learning formulations with safety guarantees for vehicle motion planning and obstacle avoidance in unstructured dynamic urban traffic.
Modeling crowd interactions and social dynamics using Spatio-Temporal Graph Convolutions (ST-GCNN) paired with Conditional VAEs and Gaussian Mixture Model latent spaces.
Real-time multi-class object detection (YOLOv11), Unpaired Image-to-Image translation with CycleGAN, and multimodal generative Text-to-Video diffusion pipelines.
High-accuracy diagnostic classification on histopathological biopsies using Vision Transformers (ViT) and Deep Residual Networks (ResNet) achieving >99.9% clinical accuracy.
Fully documented, reproducible open-source implementations, benchmarks, and quantitative reports in PyTorch.
Reinforcement Learning agent (DQN) trained to dynamically inject prompt modifiers into a frozen LLM (TinyLlama-1.1B) on the BoolQ dataset, outperforming zero-shot and Chain-of-Thought while minimizing answer length.
Physics-based multi-body modeling, Lagrangian dynamic formulation, and computed torque trajectory tracking control for a 4-DOF SCARA industrial manipulator in MATLAB/Simulink Simscape.
Multi-agent trajectory forecasting using Spatio-Temporal Graph Convolutions and CVAE-GMM multimodal latent space sampling for anticipating diverse human navigation choices in autonomous driving.
Industrial safety computer vision system featuring fine-tuned YOLOv11 for multi-class worker PPE compliance detection combined with a text-to-video generative pipeline for synthetic site monitoring.
Bridging classical tabular methods to continuous state spaces via Linear Semi-Gradient TD(0), Tile Coding, and Deep Q-Networks (DQN) with experience replay and target networks.
Unified integration of direct reinforcement learning and internal environment model simulation across stationary and dynamically altering blocking/shortcut maze environments.
Comprehensive empirical evaluation of ε-Greedy, Upper Confidence Bound (UCB), and Stochastic Gradient exploration strategies on custom obstacle GridWorld MDPs.
Exact Markov Decision Process resolution via Policy Iteration, Policy Evaluation, and Value Iteration for Poisson-distributed resource allocation and inventory management.
Deep generative image-to-image translation pipeline implementing dual PatchGAN discriminators with ResNet-based cycle-consistency and identity loss constraints on Horse2Zebra.
Clinical digital pathology diagnostic framework benchmarking Vision Transformers (ViT-B/16) against deep residual networks (ResNet-50) for fine-grained tumor subtyping.
Edge-deployable lightweight Convolutional Neural Network architecture engineered for real-time edge classification of properly masked, incorrectly masked, and unmasked individuals.
Complete implementation and comparative study of On-Policy SARSA, Off-Policy Q-Learning, Expected SARSA, and Monte Carlo Control in Cliff Walking and GridWorld domains.
Comparative study analyzing constant step-size parameter (exponential recency-weighted average) versus standard sample-average tracking in non-stationary reward distributions.
End-to-end natural language processing pipeline implementing TF-IDF vectorization, dense embedding representations, and ensemble classifiers for high-throughput customer review sentiment analysis.
Imbalanced tabular classification pipeline utilizing SMOTE oversampling, Cost-Sensitive XGBoost, and precision-recall threshold optimization across 284k transactions.
Predictive healthcare actuarial modeling using Polynomial Ridge Regression, Random Forests, and Gradient Boosting with SHAP value interpretability for demographic risk stratification.
Proficiency in deep learning libraries, robotics tools, software engineering, and mathematical foundations.
Academic teaching assistantships and machine learning engineering experience.
I am always open to discussing novel research directions, exchanging ideas on Autonomous Systems & Deep Learning, and actively exploring prospective Ph.D. positions and academic collaboration opportunities.