Zacharia Hammad

• Owned computer-vision product development across data sourcing, labeling, model training, and the user-facing application.
• Implemented shared TypeScript tool execution for VAPI and OpenAI Realtime voice agents, tying actions to verified caller identities.
• Built Neo4j identity matching and conversation context with access controls for AI applications.
• Developed YOLO/ONNX training and release tooling with MLflow tracking, evaluation, and model promotion and rollback; integrated NVIDIA DeepStream inference.
• Implemented event delivery between services using PostgreSQL and Kafka.

• Automated drone reports with Python and APIs, reducing preparation time from two hours to under 30 minutes.

• Improved submarine simulation algorithms and sensor models; documented gas-sensor workflows.

• Analyzed materials experiments with Python, MATLAB, and statistics to identify process improvements.

RepoToire

Rust and Quint

Developed a Rust runtime prototype for AI coding agents with validated execution plans, staged execution, bounded retries, and Quint-to-Rust conformance tests.

ML Pothole Detection System

Drexel Senior Design

Developed YOLO edge inference with GPS-tagged capture, backend APIs, PostgreSQL, and a geospatial dashboard for a Drexel Senior Design Championship-winning team project.

ASIC Design Optimization

Drexel Course Project

Adapted Tcl scripts for Synopsys design flows in a two-person course project, comparing timing and area across five benchmark circuits using SkyWater and ASAP7 libraries.