Muhammad Eman A.

I'm an AI engineer and QA engineer. I build AI agents, LLM apps, and automations, and I test software so it works the way it should before it reaches users. Over the past year I've shipped production AI systems end to end: multi-agent pipelines on the Claude and OpenAI APIs, full-stack apps in React and Next.js, and Python automation. I focus on systems that hold up in production: fast, reliable, and cost-aware. On a recent project I cut LLM costs 50 to 80 percent with prompt caching while improving response time. Alongside that, I work as an AI QA Engineer on a contract team, where I own functional and regression testing for an enterprise platform ahead of each release. I reproduce every bug before I report it, write clear reports with steps and recordings, and verify each fix before it ships. I also test AI features, which don't give the same answer twice, using fixed test sets and clear scoring rules so the results stay repeatable. What I can build or test for you: - AI agents and agentic workflows: agents that use tools, run on a schedule, and deliver to Slack, Telegram, email, or Notion - Tool-connected agents: AI that reads and writes to your real tools (Linear, GitHub, Google Drive, databases) through MCP and APIs - Assistant bots with long-term memory, so your agents carry context across sessions instead of starting cold - LLM apps and chatbots on the Claude or OpenAI APIs, with prompt engineering and RAG - Workflow and data automation: API integrations, web scraping, and pipelines that run on their own - Full-stack web apps: React, Next.js, TypeScript, Node.js, Supabase, and Convex - QA and testing: manual regression and exploratory testing, clear bug reports, AI output evaluation, and automated tests with pytest and Selenium I've built all of these in production, including Pulse, my open-source memory engine for AI agents. How I work: clear communication, small reviewable steps, and an early heads-up if anything is off track. I'm comfortable owning a project from first call to deployment. If you need an AI system built or software tested properly, send a message with what you're trying to solve and I'll tell you how I'd approach it.