Christian Gorton

• Compute-only IZB shadow pipeline: Implemented a novel compute-only variant of the Irregular Z-Buffer shadow algorithm, optimized from 15ms to 500µs on a 3070 Ti across four passes (GPU vertex cache, cooperative LDS loading, cell-span tuning with LDS overflow, conservative shadow map early-out). Full writeup: Reinventing the Square Wheel.
• Main opaque pass optimization: Profiled a memory-bound main render pass with RenderDoc, Nsight, and Tracy GPU zones; restructured GPU-side data layouts for cache coherency and bandwidth, reducing pass time from 5ms to 500µs.
• GPU-driven rendering: Compute-shader frustum culling with multi-draw indirect dispatch and GPU-computed draw commands. Bindless descriptor indexing for textures and buffers eliminates per-draw descriptor set updates.
• Memory and precision: VMA-managed allocation with fine-tuned buffer packing and memory type selection, FP16 mixed-precision render targets, KTX2 supercompressed textures with GPU-side format transcoding.

• Core algorithm: Each camera casts rays weighted by per-pixel inter-frame differences into a shared voxel grid; multi-view aggregation across calibrated cameras yields 3D position estimates of moving objects.
• Distributed architecture: Modular Rust workspace spanning camera units, an aggregating server, a web-based 3D client, and a Bevy-based simulator for end-to-end testing.
• Edge deployment: Targeted for deployment on Kendryte K230 RISC-V SBCs for on-device per-camera processing.

• API modernization: Contributed to migrating legacy MuleSoft APIs to Java Spring Boot, improving observability and reducing maintenance burden across backend payment services.
• Observability infrastructure: Built Dynatrace dashboards tracking production health and user flows, enabling proactive identification of latency regressions and failure patterns.
• Incident response: Led root cause analysis on recurring production issues, implementing fixes at the source rather than patching symptoms.
• Internal tooling: Built administrative tooling for production data operations, replacing manual support ticket resolution for common issues.

• ETL pipelines: Built Python and PostgreSQL data pipelines for multi-source agricultural data ingestion, with automated validation and cleaning.
• Data modeling: Designed PostgreSQL schemas and query optimization for data warehousing and business intelligence integration.
• Application refactor: Executed a performance-focused refactor of a large legacy application, introducing modular architecture and optimized database queries.
• Web applications: Delivered web applications and user management systems for enterprise clients, including a platform serving 5,000+ users.

Linear Algebra, Abstract Algebra, Analysis 1–2, Probability Theory, Machine Learning, Data Structures, Principles of Programming Languages, Calculus 1–3