Mirza O.

Your leads come in from six places. Your CRM only knows about two of them. That gap is where deals quietly die and that's the thing I fix. I'm an automation engineer. I build the plumbing between the tools you already pay for: CRM, ad platforms, phone system, email, WhatsApp, booking pages, your website. When the handoffs run themselves, your team stops doing data entry and starts closing. What I build ✅ Python-first backend stack I build with Python and FastAPI every day, backed by PostgreSQL, Docker, async workers, and clean typed APIs. Data pipelines, LLM services, embeddings and retrieval layers, all written to run in production and tested by hand before anything ships. ✅ AI-native development stack I build with Claude Code, Cursor, Codex, Antigravity, Windsurf, and Lovable every day plus Claude, GPT, and Gemini APIs under the hood of what I ship. Rapid prototypes, agentic coding workflows, and full production builds, all reviewed and tested by hand before anything goes live ✅ Lead pipelines end to end Meta and LinkedIn Lead Ads, website forms, and booking pages deduped, enriched, and routed straight into HubSpot, GoHighLevel, or Airtable with full UTM attribution intact ✅ AI SDR and outreach systems Prospect research, personalized first touch, follow-up sequencing, reply handling, everything logged back to the CRM ✅ AI voice agents and chatbots Vapi, ElevenLabs, and Retell for bookings, intake, qualification, and 24/7 front desk coverage — with call transcripts and outcomes written back to your CRM automatically ✅ Twilio SMS, voice, and messaging systems Programmable Voice and SMS, IVR menus, call routing and forwarding, two-way SMS conversations, appointment reminders, OTP verification, and A2P 10DLC registration handled so your messages actually deliver ✅ WhatsApp Business Cloud API Multi-number inbox, template approval and fallback, 24-hour window handling, media messaging, and campaign sending — built on Meta's Cloud API directly, not a reseller wrapper ✅ RAG assistants Your site, docs, and PDFs turned into an assistant that answers using your own content (Supabase pgvector, Pinecone) ✅ n8n, Make, and custom Python workflows Webhooks, retries, idempotency, error alerting. Built to survive a bad API day, not just a demo call ✅ Phone system to CRM 3CX call journaling, click-to-call, screen pop on inbound ✅ Internal dashboards Next.js and Supabase, so you can actually see what the automation is doing Where it saves you the most money: ✅ MVPs and internal tools built in days, not months ✅ Legacy codebases documented and refactored instead of rewritten ✅ Repo-wide changes done consistently instead of file by file ✅ If you're already using Claude Code or Cursor in-house, I can set up your CLAUDE.md, project rules, and agent workflows so your own team ships faster too Recent build (B2B tech client, ongoing) An AI SDR and inbound revenue system: FastAPI backend with Celery workers on DigitalOcean, two-way HubSpot sync, LinkedIn and Meta lead webhooks, a WhatsApp campaign module, multi-inbox email sending, Cal booking capture with 10-field UTM tracking, and a live inbound dashboard. Ad click to conversation to CRM deal, with nothing typed by hand. Also strong on the web side, when your automation needs a front end: Next.js and React apps, custom WordPress themes and plugins, WooCommerce, REST API work, Core Web Vitals, multilingual and SEO-ready builds. How I work Small scope first, working system fast, then expand. Every delivery comes with documentation so your team isn't stuck depending on me. I overlap with US business hours. Start here: tell me the one process that eats the most hours in your week. I'll tell you honestly whether automating it is worth the money and if it isn't, I'll say so.