Angel David L.

I build the hard parts of web applications. The parts most developers avoid. For the past 5 years I've worked across the full stack with JavaScript, TypeScript and Python, taking problems that start as "how do we even make this work?" and turning them into systems running in production. Lately that's been AI systems and AI agents. Before that, it was 3D in the browser. Same engineering approach, different terrain. AI systems I've built: - Ailyus: an AI support automation platform I built solo. It fixes the problems that break businesses day to day, like failed Salesforce or Shopify integrations or stuck user accounts, by letting you program AI-governed flows that diagnose and resolve them. Conversation engine, a connector system that turns any API into MCP tools, playbooks, and approval flows. - theORQL: think Cursor or Claude Code, but for debugging the frontend. A VSCode extension whose AI agent connects to Chrome DevTools Protocol, so it can actually see the DOM and runtime state of your app instead of guessing from the code. - Luna Sentinel: a node-based AI automation system built on LangGraph, with PostgreSQL + pgvector for RAG, Redis, and a multi-provider LLM router. It monitors job boards, scores every posting with LLMs against a profile, and surfaces the ones worth pursuing. Flows are defined as JSON and executed as graphs. Building these took me through the full AI engineering stack: MCP, context management, RAG and embeddings, prompt engineering, node-based AI workflows, OpenAI and Anthropic APIs, and token cost optimization. When your posting asks for an AI MVP, an AI agent, or an LLM integration, I've already done the learning on real production systems, not on your budget. 3D and frontend: I also build interactive 3D for the browser: a real-time skateboard configurator for e-commerce (Three.js / React Three Fiber), audio-reactive GLSL shaders on a production site, scroll-driven experiences with GSAP and Motion. My frontend work is engineering-first: React, Next.js, TypeScript, UI performance, and faithful implementation of your designs down to the pixel. Backend: Node.js, NestJS, Python. PostgreSQL, MongoDB, Supabase, Redis. How I work: I break your problem into parts, propose an architecture, and tell you honestly what's worth building and what isn't. Clients delegate development to me so they can focus on the business. If you have an existing team or codebase, I plug into it. Not every project is complex, and that's fine. If you need a clean website or a straightforward dashboard, you get the same engineering: shipped fast, built to last, no surprises. And if your project does have a hard part, that's the part I enjoy most. P.S. What's the piece of your project you suspect will be the most difficult? Start there with me.