Muneeb S.

𝗜 𝗯𝘂𝗶𝗹𝗱 𝗔𝗜 𝗮𝗴𝗲𝗻𝘁𝘀 𝘁𝗵𝗮𝘁 𝗿𝘂𝗻 𝗶𝗻 𝗽𝗿𝗼𝗱𝘂𝗰𝘁𝗶𝗼𝗻 — 𝗻𝗼𝘁 𝗷𝘂𝘀𝘁 𝗱𝗲𝗺𝗼𝘀. 5+ years as a software engineer, the last year+ focused on deploying LLM-powered agents that actually run unattended in real business environments. 𝗔𝗜 𝗔𝗚𝗘𝗡𝗧𝗦 & 𝗢𝗣𝗘𝗡𝗖𝗟𝗔𝗪 — 𝗦𝗣𝗘𝗖𝗜𝗔𝗟𝗜𝗦𝗧 𝗘𝗗𝗚𝗘 I deploy, configure, and run OpenClaw agents in real business environments — not just wire up an API key and call it done. ➞ Built and maintain a production OpenClaw email agent for a networking/IT services company: reads Outlook/Exchange via Microsoft Graph, categorizes every inbound email into 16+ workflow categories, drafts replies, and posts real-time briefings to Microsoft Teams via the Bot Framework — running unattended on a VPS with systemd, cron scheduling, and proper credential rotation ➞ Hard safety rules by design: agent never sends email or touches calendars without explicit confirmation — drafts only, human approves ➞ Comfortable working directly in SOUL.md / agent identity & workflow files, not just surface-level config 𝗦𝗘𝗠𝗔𝗡𝗧𝗜𝗖 𝗦𝗘𝗔𝗥𝗖𝗛 & 𝗥𝗔𝗚 — 𝗔𝗖𝗧𝗜𝗩𝗘𝗟𝗬 𝗘𝗫𝗣𝗔𝗡𝗗𝗜𝗡𝗚 Built embedding-based semantic search (cosine similarity matching over a pre-built vector index) for a healthcare client's Firebase backend, using the Vercel AI SDK and Gemini. Currently extending this into full RAG pipelines (chunking, vector DB, grounded generation) — happy to be upfront that this is a growth area, and I bring real engineering rigor to it rather than a copy-pasted LangChain tutorial. 𝗕𝗔𝗖𝗞𝗘𝗡𝗗 & 𝗔𝗜 𝗜𝗡𝗧𝗘𝗚𝗥𝗔𝗧𝗜𝗢𝗡 ✅ Node.js, TypeScript, Python, Express ✅ Firebase Cloud Functions, PostgreSQL, JWT auth ✅ REST API design, webhook-based integrations ✅ OpenAI API, Gemini, Claude API ✅ Multi-tenant SaaS architecture (currently running a GDPR-compliant, multi-tenant healthcare platform on ERPNext/Frappe) 𝗠𝗢𝗕𝗜𝗟𝗘 & 𝗙𝗨𝗟𝗟-𝗦𝗧𝗔𝗖𝗞 (𝗪𝗛𝗘𝗥𝗘 𝗜𝗧'𝗦 𝗨𝗦𝗘𝗙𝗨𝗟) Flutter development when the right answer is a custom app rather than an automation — I'll tell you honestly which one your problem actually needs. 𝗜𝗡𝗗𝗨𝗦𝗧𝗥𝗜𝗘𝗦 Healthcare, IT/Networking, Government/Public Sector, Consumer Apps 𝗛𝗢𝗪 𝗜 𝗪𝗢𝗥𝗞 1. Understand the actual problem before proposing a solution — sometimes AI isn't the right call, and I'll say so 2. Architect for production from day one — not a demo you can't scale 3. Build fast, test against real data 4. Document and hand over cleanly (I write real handover docs — configs, credentials, rotation schedules, the works) Send me your use case — I'll give you a straight answer on what it actually takes.