Inventory
Production-grade tech stack
A selective list of the tools I use to solve real-world problems. Focusing on performance, maintainability, and measurable impact.
AI/ML & Automation
Core LLM integration. Prefer Claude 3.5 Sonnet for complex reasoning/coding and GPT-4o for high-throughput tasks.
Orchestrating stateful, multi-agent workflows with cycles—crucial for non-linear reasoning chains.
Preferred low-code automation engine for rapid prototyping and connecting diverse legacy APIs with LLM nodes.
Language of choice for ML research, data mining, and building specialized agents with core science libraries.
Enabling RAG architectures for context-aware agents with persistent semantic memory.
Backend & Infrastructure
Server Components significantly reduce client-side JS. Using for this portfolio for SSR/SSG performance.
Scalable backend services with TypeScript. NestJS for enterprise-grade dependency injection and architecture.
Production-ready hosting with built-in CI/CD, staging environments, and global scale on the Free Tier.
Type-safe DB access and reliable relational storage for enterprise data consistency.
Used for performance-critical components (like the Torven core) and high-concurrency tooling.
Frontend & Design System
Industry standard for building robust, type-safe interactive interfaces.
Zero-runtime CSS for maximum performance and 100/100 Lighthouse scores. Avoiding Tailwind to maintain full control over the design system.
Managing this monorepo to enable shared UI components and config across multiple apps.