• Developed PBPK models using MATLAB and SimBiology with mass-balanced ODEs to predict drug ADME and optimize nanoparticle (NP) therapeutic delivery systems, calibrating parameters with in vitro and in vivo pharmacology data.• Created mechanistic PK/PD models to simulate disease progression under immunotherapy, chemotherapy, and micro-RNA treatments, identifying 3 synergistic drug combinations and optimal dosing schedules for NSCLC.• Conducted local and global sensitivity analyses, pinpointing key PBPK model parameters affecting drug biodistribution.• Generated virtual populations via Latin Hypercube Sampling and Monte Carlo simulations to improve population-scale PK predictions.• Translated preclinical PK data to human parameters using allometric scaling, evaluating clinical endpoints (e.g., PFS, OS), and optimizing dose regimens to enhance therapeutic efficacy.• Utilized nonlinear mixed-effects modeling (popPK) with NONMEM to analyze inter-patient variability in drug concentrations from clinical data, informing clinical decision-making.• Trailblazed ML integrated PBPK modeling, successfully tuning and evaluating 75+ ML and DL models using SciKit-Learn, TensorFlow, and PyTorch to classify organ-specific toxicity and predict bioaccumulation in vitro and in vivo.• Pioneered generalized therapeutic NP design guidelines by applying model explainability techniques (SHAP, LIME, PDP), quantifying the structure-activity relationships of 7 key factors influencing organ-specific cytotoxicity mechanisms.• Proven expertise in handling discrete observations, continuous measurements, and time-series data types.• Presented 20+ invited talks and poster sessions at national and international conferences, securing 8 awards, demonstrating exceptional communication and presentation skills.

• Spearheaded the development of Graph Convolutional Networks (GCN) using Keras and TensorFlow for classifying lymphatic tissue micro-environments from 3D cell surface data, achieving ~85% accuracy.• Managed diverse data types, including Whole Slide Images (WSIs) and genotype data, to classify cell phenotypes and infer gene expression levels, showcasing expertise in data handling and interdisciplinary analysis.• Implemented dimensionality reduction techniques, specifically variational autoencoders (VAEs), to handle high-resolution WSI data, significantly enhancing computational efficiency and model performance.• Enhanced dataset size by 50% through synthetic data generation with VAEs (via learning latent space representations) and generative adversarial networks (GANs), facilitating more robust model predictions.• Optimized GCN by leveraging transfer learning for performance improvements and benchmarked against existing computational models (‘GraphSAGE,’ ‘attri2vec’).• Collaborated across 4 different labs, fostering a multidisciplinary approach to development.

- Systematically researched the sustainability of FMCW RADAR and optical automotive perception sensors.- Analyzed and elected PCB materials to prolong lifespan and wear-out for next-generation FMCW RADAR sensors.- Investigated and refined current manufacturing methods utilized to mitigate the causes of sensor ageing and malfunctioning.- Established a framework for sensor recycling strategies relative to fabrication techniques, international WEEE policies, market demand, and structural conditions.- Defended original research via an oral presentation by deliberating the environmental, technical, and economic benefits of material reuse.

• Investigated mathematical models to assess the impact of confining halide perovskites in Single-Walled Carbon Nanotubes (SWCNTs) on electrical properties and bandgap, demonstrating advanced computational and analytical skills.• Conducted XRD on 30+ samples to assess crystalline structures and measure bandgap modifications for model validation.• Applied advanced microscopy techniques (SEM/TEM) to characterize SWCNT-perovskite composites and EDS for elemental composition analysis of halide perovskites within CNTs, showcasing meticulous technical expertise.• Employed MATLAB and Python for advanced image processing, including Gaussian blurring and Fast Fourier Transforms.• Utilized Materials Studio and LAMMPS for nanocrystal simulation, modeling atomic interactions to predict electronic properties influenced by SWCNT characteristics, demonstrating proficiency in grasping new concepts and ideas quickly.

• Redesigned a Supervisory Control and Data Acquisition (SCADA) GUI system for AFP detectors within the ATLAS accelerator using Linux, the Joint Controls Project (JCOP) framework, and Siemens WinCC OA tools.• Significantly reduced data latency and enhanced the ATLAS Forward Proton detector's spatial and temperature sensor feedback accuracy through programming specialized functions using C.• Installed and configured the SCADA interface for operators in the ATLAS control room, ensuring seamless integration with existing systems while meticulously documenting the development processes to facilitate reproducibility.• Demonstrated capacity to work collaboratively and communicate effectively in a demanding international research environment, culminating in an award-winning project.• Integral part of daily operations and is actively used by technicians and operators in the particle accelerator's control room, demonstrating sustained impact on high-energy physics research.

• Awarded a Nuffield Foundation Fellowship for research at Bluefruit Software, focusing on medical image processing efficiency.• Developed a memory-efficient compressed file system for Philips ultrasound systems using C on Linux, improving storage and retrieval processes for high-resolution images.• Implemented Huffman coding and utilized cyclomatic complexity equations to optimize data compression without compromising image quality and system efficiency.• Employed C/C++, Python, ‘Zlib’/’Gzip’ for compression, and ‘OpenSSL’ for data security.• Integrated POSIX Threads for enhanced multi-threading capabilities and conducted comprehensive benchmarking with ‘iozone’ and ‘fio’.• Utilized Docker to simulate varied operating conditions and executed extensive stability, load, and scalability tests to ensure reliability and performance under real-world conditions.• Collaborated with hardware engineers to ensure compatibility with Philips’ proprietary ultrasound hardware and managed software development stages using Git for version control, ensuring team collaboration and consistent updates.