Muhammad Zia

WORK HISTORY

EDUCATION

ABOUT MUHAMMAD ZIA

Senior Full Stack Engineer based in London, building production systems where AI is a first-class component — not a tacked-on demo. Over the past 5+ years I\'ve led technical direction on products used by 200+ international teams processing 10M+ records. I now work at the intersection of full-stack engineering and applied AI: designing agentic workflows, integrating LLMs into real products, and engineering the connective tissue that turns model outputs into reliable user experiences. What I ship: → Built an MCP (Model Context Protocol) server at Convert, exposing the experimentation platform to LLM clients as first-class tools — letting agentic workflows query experiments, read results, and trigger actions against the platform using Anthropic\'s open standard → Shipped an AI-powered email editor at Instantly that detects spam-trigger language in real time and highlights it inline — directly improving inbox deliverability for high-volume outbound users → Built natural-language lead search at Instantly — users describe their ideal customer profile in plain English and the system returns the best-fit leads, replacing rigid filter UIs with semantic retrieval→ Architected three-layer error-handling system at Instantly (React boundaries, API interceptors, global runtime handlers) — 99.8% capture rate, 47% fewer support tickets → Shipped Multi-Arm Bandit optimization and core Visual Editor features at Convert, contributing to 34% conversion lift across enterprise experimentation clients → 3x\'d developer velocity at Pocketnote through design system + CI/CD overhaul — 68% faster deployments,$2.4K/month cloud savings→ Currently owning end-to-end architecture on an industrial schematics platform used across global marketing and operations (refineries, manufacturing, logistics) Stack I go deep on- Full-stack: React, React Native, Next.js, Node.js, TypeScript, GraphQL, MongoDB, REST, Stripe, Paddle and many more- AI integration: OpenAI, Anthropic Claude, LangChain/LangGraph, RAG pipelines, vector stores, prompt engineering, human-in-the-loop UX, streaming/async UIs, agent observability - AI-driven development: Claude Code, Cursor, Copilot — used daily to compress delivery cycles without sacrificing code quality What I bring beyond code: architectural judgment, clean abstractions, and the discipline to design AI features that degrade gracefully when models misbehave — rate limiting, guardrails, PII masking, fallback paths, error recovery. m••••••••@icloud.com github.com/mughalhere mozia.dev