Md Navid Akbar, Phd Email & Phone Number

Fullerton, Ca, Us

Fullerton, Ca, Us

Woodbury, Minnesota, Us

• Measure ST segment changes by segmenting multi-channel cardiac time series (6 ms, validation), to detect myocardial ischemia with filtering and signal processing, for the wearable Vios patient monitor [MATLAB]• Automate data quality and failure mode analysis, test hardware and analyze collected data, and prepare external datasets for validating algorithm to comply with applicable ANSI/AAMI industry standards• Develop a U-Net based, ECG segmentation DL model PoC for in-progress, server-side ML workflow [Python]• Derive algorithm features and design choices by adhering to clinically derived user inputs, communication across cross-functional teams (HW, SW, system, UI, marketing), and FDA 510k regulatory requirements

Andover, Massachusetts, Us

• Tracked in real-time (0.5s latency) medical devices under low-field MRI, with a hybrid CV instance segmentation (YOLO) and motion kinematics model, improving detection over SOTA (DINO, SAM) [OpenCV]• Created image annotations (100+), and developed basic UI for region-of-interest selection in video inference• Authored an NSF grant pitch, and maintained documentation following FDA guidelines

Boston, Ma, Us

• Modeled 3D functional neuroimaging by brain stimulation (TMS) to causal EMG response via deep learning convolutional regression (16% error reduction) with interpretable domain knowledge fusion [TensorFlow]• Designed a first-of-its-kind VAE inspired, generative 3D convnet for inverse imaging of brain motor excitation mapping from TMS-induced 1D EMG muscle activity (R2: 99.6% max), for rehabilitation planning [TensorFlow]• Designed novel multimodal fusion algorithms (2) with explainable ML, for seizure classification and biomarker identification (2 from 564), with unsupervised treatment of missing data in dMRI, fMRI, EEG [Scikit-learn]• Classified binary seizures (8% AUC increase) using brain dMRI preprocessing, feature selection, boosted trees, and statistical analysis [Scikit-learn, Pandas] • Preprocessed and set up (2) in-house neuroimaging and neuromodulation data pipelines [FSL, MATLAB]• Analyzed a large data set (500k words) using an NLP model on a distributed computing backend [PySpark]• Improved (8% IOU) a crowdsourced image annotation task using HCI theories [HTML, AWS S3]• Revamped an Android app (phone, watch) to log wellness and IMU sensor activity (10+ users) [Java]• Assisted an NIH grant renewal preparation, and mentored graduate data science courses (35+ students)• Received two best paper awards (IEEE and ACM conferences) and funding (Fellowship, NSF travel grant, etc.) • Reviewed journals and conferences (10 papers), and organized technical seminars in Dept. of ECE

Amsterdam, Noord-Holland, Nl

• Predicted image-guided, surgical device trajectory with multimodal fluoroscopy and time series robotics data using a Conv-LSTM recurrent neural network, reducing (up to 85%) radiation exposure [PyTorch]• Quantified lung disease severity from chest x-ray clinical imaging using Siamese convolutional neural network with DenseNet, improving (4% AUC) over SOTA with smaller data, without expensive pre-training [PyTorch]• Developed a self-contained workflow for ML deployment [Docker]

Richardson, Texas, Us

• Developed two signal processing algorithms: improved adaptive wireless receiver performance with lower complexity (3x savings) | improved (28dB) RSSI of indoor 5G reception by beam steering [MATLAB]• Denoised and classified (90% accuracy) ECG signals with wavelets, SVM, and cloud [MATLAB] • Drafted a research review (22 papers) on noninvasive glucose and alcohol wearable sensor monitoring • Taught embedded systems lab using TI MSP 430, focusing on programming, I/O, and sensor analytics [C] • Assisted preparation of grant proposals (NSF, DoD) and teaching engineering courses [400+ students]

Tuskegee, Alabama, Us

• Mentored an undergraduate engineering lab [10+ students]

Shenzhen, Guangdong, Cn

• Designed optical, FTTX, IP core, and microwave communication product solutions (4 telecom clients)

Doctor Of Philosophy - Phd, Computer Engineering

Master'S Degree, Electrical, Electronics And Communications Engineering

Bachelor Of Science (B.Sc.), Electrical And Electronics Engineering