Projects

Sepsis Global Attribute Atlas (Clinical NLP Hackathon)

Built a high-performance Streamlit dashboard to automate the extraction of clinical variables (Biomarkers, Severity Scores, Treatments) from unstructured medical literature. Implemented a metadata-aware parsing engine to ensure deep traceability back to exact page numbers and evidence context.

• Engineered an Attribute Consensus Matrix and Network Relationship Graph to visualize variable co-occurrence across 30+ papers.
• Integrated Weaviate Vector Database via Docker Compose for robust document indexing and retrieval.
• Developed a "Consensus Cloud" visualization to identify high-importance variables for clinical benchmarking.

MineSight: 3D Conveyor Path Intelligence (AMT Hackathon)

Architected an interactive 3D layout editor for mine conveyor planning, bridging the gap between graph-theory simulations and spatial engineering. Developed a custom synchronization engine to align multi-strand conveyor paths with high-resolution digital terrain models.

• Implemented a Terrain-Aware FIT Engine with automated Z-clearance logic to snap conveyor nodes to wireframe displacements while maintaining structural safety gaps.
• Engineered a unified data ingestion pipeline supporting NetworkX (Node-Link JSON), DXF (CAD), and XLSX/CSV for seamless engineering-to-visualization workflows.
• Developed a real-time 3D/2D Dual-Viewport system using Three.js (React Three Fiber) to visualize belt tension, elevation profiles, and path safety warnings.

Integrating Traffic Light Status for MPC Planner in ACDC Suite

Developed the existing Model Predictive Control Planner incorporated by Automated and Connected Driving Challenges ACDC Research Project with C++, ROS to integrate traffic light status into the MPC planner. This involved modifying the planner to account for traffic light states using SPAT and MAP messages.

Integrated Maintenance Solution — Fraunhofer ICNAP Hackathon

Architected an end-to-end industrial IoT solution to bridge sensor telemetry with mixed reality maintenance tasks.

• Developed "Data Kraken" and "Data Stream" to unify multi-protocol inputs (OPC-UA, MQTT, REST API).
• Explored Unity integration to overlay real-time sensor data onto Mixed Reality environments for technicians.

Customer Experience Enhancement with Generative AI

Developed a web application using LLMs and a custom data corpus to generate context-aware, personality-driven responses for GitHub issue management.