Jade Glass

• Built an automated device-characterization and root-cause-analysis pipeline that retrieves wafer data from GlobalFoundries’ Redshift manufacturing database, analyzes device behavior in Python, and generates 1,420 standardized RCA packets across 1,327 lots. Developed a permission-gated AI agent that compares results across wafers, explains findings in plain language, and reduces investigations that could take days to immediate engineering insights.

• Developed CauseLens, an AI-assisted anomaly detection tool processing 500,000 rows of semiconductor manufacturing data in about 3 minutes, replacing hours of manual SHINY chart review and enabling quicker decision-making. Secured approval from SHINY platform team to integrate its outlier-ranking framework as a future native feature across Fab 8.

• Created HoldsWatch, an automated daily hold-management system that prioritizes high-severity lots, assigns owners, records dispositions and action items, and posts meeting outcomes to Microsoft Teams via SharePoint and Power Automate, enhancing management of up to 47 held lots daily.

• Led data-driven process optimization to reduce residual solvents in flame-resistant Nomex fibers, enhancing long-term safety for customers exposed to the material in extreme heat environments.
• Eliminated the need for multi-million-dollar equipment upgrades, saving the company approximately $7M through targeted experimentation instead of full system replacement.
• Utilized Minitab and RYS for experimental design and analysis, facilitating communication between operations, R&D, and management.
• Organized and submitted experimental campaign data, streamlining R&D workflows.

• Developed a Python-based algorithm to digitally map the township's entire sewer infrastructure, previously undocumented, by cross referencing 70+ years of blueprints and mismatched datasets.
• The tool modeled flow direction, ranked manhole connections, and calculated upstream aggregates, supporting accurate emergency cutoff operations and saving the township ~$50,000 in external consulting costs.
• Mapped sewer network flows from over 10,000 residential and commercial sources to two downstream processing sites, improving data integration.
• Built custom logic-based parser to reconcile misaligned IDs across legacy maps and spreadsheets, validating outputs with Python plotting libraries for accuracy.
• Designed lead testing protocol for 50,000-ft² school facility, ensuring compliance with regulatory standards.
• Authored a sediment flushing and sampling sequence aligned with EPA guidance.
• The plan was adopted by an external testing lab and successfully cleared the school for reopening.