Hamza Iqbal

Graduate researcher from November 2025; research associate since completing the M.S. Advisor: Nathan Jacobs.

• Co-authored Tessellating the Earth (ECCV 2026): a location encoder built from learnable spherical Voronoi partitions with shared global semantic tokens; state of the art across geospatial classification and regression benchmarks.
• Lead a project on efficient retrieval for planet-scale image geolocalization: hierarchical search over a multiresolution gallery that matches exhaustive retrieval at a fraction of the encoding cost and extends to object-level queries. Paper under review.
• Train contrastive CLIP-style vision transformers on large-scale multispectral satellite imagery; run distributed experiments on SLURM clusters with PyTorch Lightning; build cross-modal (text-to-image, image-to-image) retrieval pipelines over hierarchical geospatial tiles.

• Lead the data and AI practice; designed and built the analytics platform behind research-strategy diagnostics and innovation scans for research universities.
• Built collectors and a data lake over NSF HERD, OpenAlex, USPTO, NIH RePORTER and AUTM data; peer benchmarking of research expenditures, output, funding mix and technology transfer.
• Topic clustering of institutional research portfolios with SPECTER2 embeddings, BERTopic and Leiden community detection; automated LaTeX and PowerPoint report builders.

• Wrote and owned PySCOTT, a link-budget and scan-configuration modeling tool (Python, Dash) predicting SNR across range, Doppler and angle from simulated or measured antenna patterns; its curves set production test limits and drove customer coverage maps.
• Developed automated test and validation pipelines with 2D/3D coverage visualizations for sensor characterization across hardware configurations: turntables, interference simulation, close-range experiments.
• Modeled link budgets for URA/ULA and sparse-array sensor configurations; built analysis tooling for the radar DSP pipeline to diagnose noise floor, signal magnitude, vector misalignment and phase mismatch.
• Applied ML to classify false alarms, multipath and interference in real-time object detections; built the CLI used company-wide for bench tests and deployed automated software testing that caught issues during active development.

• Converted HL7v2.x lab results to FHIR and integrated them with PracticeFusion EMR records, storing unified patient data in PostgreSQL and MongoDB for downstream analytics.
• Built Spark ETL pipelines consolidating lab results, billing and patient records from heterogeneous clinical sources into a queryable data lake, visualized in Apache Superset; worked within HIPAA constraints.

• Managed clinical datasets for fall-prevention and Alzheimer’s studies with the Knight ADRC; cleaned and structured 5,000+ participant records, including unstructured clinical notes, for cohort selection.
• Statistical analyses of health-outcome and time-series data (Python, MATLAB, SPSS); data dictionaries and merges across instruments and study arms; REDCap filters and conditionals to prevent recurring data-entry errors.

Coursework: Bayesian machine learning, large language models, advanced computer vision, deep reinforcement learning, text mining, data mining, information theory, rapid prototyping.

Teaching assistant, CSE 4470: Automata & Theory of Computation (Spring 2026).

Statistical learning and classical ML, neural networks, computer vision, NLP and generative AI, with end-to-end projects in image classification, recommendation systems and model deployment.

GPA 3.86. Coursework: probability theory, statistical inference, algorithm design, numerical linear algebra, financial derivatives, introduction to machine learning.

• Built a low-cost, open-source control and timing system for behavioral experiments in Python and C on Arduino, with a custom PCB designed in Eagle: 85% cheaper and 87.5% lower latency than the commercial alternative.
• Frequency-domain analysis of 3+ TB of mouse EEG toward a non-invasive deep-brain-stimulation approach. IMSD fellow from August 2019.

• Fixed bugs and added features to QuB, the lab’s Delphi-based ion-channel simulation software; ran electrophysiology experiments with two-photon imaging; trained in patch clamp.
• Developed a low-cost multispectral imaging system, prototyped lab tools with CAD and 3D printing, and migrated the lab website to Google Cloud.

• Analyzed participant retention in Knight ADRC Alzheimer’s studies with SPSS, the Google Maps API and regression on a cleaned 5,000+ participant dataset; pilot results presented at IAGG 2017.