Safi Rahman

I’m a Full Stack AI Developer with 6+ years shipping production web applications, AI agent platforms, and voice AI systems for clients across the US, UK, Canada, Australia, Japan, and the Middle East. I pride myself on end-to-end delivery—frontends, backends, cloud infrastructure, and AI-powered features—so solutions actually work in production.

Recently, I architected a SaaS platform for configuring and deploying AI agents at scale, powering 50+ deployed AI employees and handling 5,000+ monthly AI-driven phone calls. I built the real-time voice infrastructure using Twilio PSTN/SIP, Deepgram STT, ElevenLabs TTS, and LiveKit streaming, with encrypted credential management and automated live call transfers.

I also focus heavily on reliable agent behavior. I engineered durable workflow orchestration with Temporal to eliminate ~40% silent agent failures and enable reliable multi-step execution across LLM providers, plus RAG-powered vector memory (Pinecone, mem0) with structured data extraction that improved response accuracy by ~35% on domain-specific queries.

Across roles, I’ve led small teams, collaborated directly with non-technical stakeholders, and reduced delivery friction by standardizing build-test-deploy methodology. From voice/chat agent platforms to secure fintech dashboards and performance-optimized UIs, my approach is consistent: use AI-assisted development tools to ship faster without compromising code quality, and measure outcomes end-to-end.