Joseph Gonzaludo

• Configured and valued European gas storage and transport structures from Endur (ETRM), modeling injection and withdrawal rights under volume and price constraints with backward induction.
• Built a plugin and MCP server that exposes Endur's valuation logic as a callable tool inside an internal agentic AI platform, so the team can query and automate storage and transport valuations in natural language.
• Extended the parameter evaluation logic of a Least Squares Monte Carlo pipeline used for stochastic optimal control in commodity derivatives pricing, supporting conditional expectation estimates for path-dependent exercise decisions.

• Engineered 5+ production features for the Goddard platform in React and Python, cutting manual review time by about 50% by optimizing distributed SQL workflows.
• Diagnosed cross-service failures by tracing malformed requests and data mismatches, then added automated validation that improved deployment stability.
• Wrote Python and SQL scripts that surface architectural defects before rollout, improving operational consistency and reducing support load for 100+ high-stakes users.

• Built a full-stack interactive AR prototype with TypeScript, WebGL and cloud deployment tooling for a Lazada feature projected to generate $23–30M in annual revenue.
• Placed 2nd of 500+ global teams in the Alibaba Global E-Commerce Challenge with an AR “view in your room” app meant to reduce purchase uncertainty in Southeast Asian e-commerce, built with React and Google's model-viewer library. Pitched live at Alibaba HQ to 70+ engineers and executives, including C-suite leaders.

• Architected core platform workflows in React and TypeScript, managing deployments and triaging UI and API issues for a large-scale rollout to 230+ users.
• Analyzed user activation and drop-off with SQL and Python, then shipped UX changes that raised platform completion by about 40%.

• Reduced runtime latency by 27% in a machine translation system by implementing optimized A* decoding and beam search pruning in Python, while maintaining high translation quality.
• Implemented IBM Model 1 with NumPy vectorization, achieving significant speedups over iterative approaches and surfacing data alignment defects through likelihood tracking.

B.S. Computer Science. Minors in Creative Writing and Entrepreneurship & Management.