I'm a Frontend Engineer who builds AI-powered products for the web and mobile.
I live at the intersection of modern web engineering and applied LLMs — designing RAG pipelines, agentic workflows, and context-engineered chat experiences that real users actually ship with. I'm not training foundation models; I'm using them to build things people love.
const syket = {
role: "Frontend Engineer · AI Product Builder",
building: ["RAG apps", "AI agents", "Chat UIs", "MCP tools"],
stack: ["Next.js", "React Native", "LangChain", "Vercel AI SDK"],
philosophy: "Ship LLMs into production — not demos",
learning: ["LangGraph", "Eval pipelines", "Agentic UX"],
funFact: "Solved 100+ DSA problems · I write on Medium"
};Building production-grade AI features — RAG, agents, and context pipelines that ship.
What I actually build:
| Area | What it means in practice |
|---|---|
| 🔍 RAG Systems | Chunking strategies, embeddings, hybrid search, re-ranking, citation UX |
| 🤖 AI Agents | Tool-use, multi-step reasoning, LangGraph state machines, human-in-the-loop |
| ✨ Prompt Engineering | Structured outputs, few-shot, chain-of-thought, output validation |
| 🧩 Context Engineering | Token budgeting, context windows, memory, eval-driven iteration |
| 💬 Streaming Chat UIs | Vercel AI SDK, RSC, generative UI, optimistic updates |
| 🔌 MCP Integrations | Model Context Protocol servers, custom tools for Claude & agents |
- 🧪 Eval-driven AI development — Braintrust, Langfuse, custom eval harnesses
- 🕸 Multi-agent orchestration — handoffs, supervisors, parallel agents with LangGraph
- 🎨 Generative UI — streaming React components from LLM tool calls
- 🔭 On-device & edge inference — running small models with WebGPU + Transformers.js
📚 Read more → medium.com/@syketb
If you're building something with AI on the web — or want to chat about RAG, agents, or shipping LLM features — my inbox is open.
"The best way to predict the future is to build it — one prompt at a time."