>*I create software that works without intervention once deployed. The architecture begins with an analysis of data flows, failure modes, and attack surfaces, not with user interface mockups.*
*I build infrastructure that operates without supervision: full-stack applications under real user load, analytical pipelines on heterogeneous data, systems designed with attack surface minimization from day one. No demos. No prototypes. Only deployed solutions.*
|Delivered|Context|Outcome|
|:-|:-|:-|
|E-commerce platform|Regulated healthcare setting|Full order lifecycle: inventory sync, payment reconciliation, sub-second storefront. Sustained 500+ orders/day during peak sales. Zero downtime.|
|Enterprise document intelligence|200000+ heterogeneous technical document|Proprietary retrieval architecture - no off-the-shelf RAG. Precision 0.92, query latency under 120 ms. Deployed as internal service|
|Сlinical decision dashboard|Regulated healthcare setting|Real-time telemetry > actionable insight. Reduced clinical triage latency by 40%. Auth model passed HIPAA audit.|
|Web-native AR visualizer|Global brand product configurator|Browser-based via WebXR. No app install. Maintains 60 FPS on mid-tier mobile GPUs. 15 000+ monthly active users.|
|Deployment automation suite|Early-stage SaaS with manual release cycle|Git push > production in 90 seconds. Eliminated 45 minutes of error-prone human steps. Zero rollbacks since adoption.|
|API threat surface reduction|Pre-launch fintech stack|Eliminated 7 critical vectors: auth bypass, IDOR, SSRF. Redesigned to remove entire exploit classes at architectural level.|
***Engineering foundation***
* *C++ memory model, RAII, zero-cost abstractions (Lippman 5e)*
* *OOP design beyond syntax, interface segregation (Zandstra)*
* *PHP internals: opcode, SAPI layer awareness (Jank)*
* *Security: attack surface reduction, not checklist compliance*
**Stack - only what’s necessary**
* *Frontend: Vue 3, React, Tailwind, Three.js/WebXR*
* *Backend: Laravel, Node.js, raw PHP when justified*
* *Big Data: custom ingestion > feature extraction > serving*
* *ML: own LLM variants, fine-tuned on domain corpora, server-deployed*
* *Data: PostgreSQL, MongoDB, time-series stores*
* *Infra: Docker, Kubernetes, AWS, GitHub Actions*
* *Security: threat modeling, hardened auth, pentest background*
* *Engine: C++, Unreal Engine for 3D logic when required*
WordPress used only if client requires self-editing - always with custom admin interface. Otherwise headless.
Live deployments: [https://existencecode.com](https://existencecode.com)
Code: [https://github.com/TaylonMcFly](https://github.com/TaylonMcFly)
Available for contracts requiring engineering depth. Direct clients only. Rate: $50/h or fixed scope.
DM with:
Problem statement
Data characteristics (if applicable)
Timeline
No agencies. No speculative work. Let’s build something that endures.