About
I build AI systems that actually run in production.
Right now I have a voice AI answering patient calls for a dental practice, booking appointments autonomously 94% of the time. Another one handling emergency dispatch for a plumbing company. And a WhatsApp AI platform where multiple businesses each get their own AI agent.
I also built an MCP server that lets Claude interpret medical lab data — pulling live results from Epic's FHIR sandbox, flagging critical values, and generating clinical summaries.
I came into this from healthcare — BSc in Medical Laboratory Science — which is probably why I keep ending up on projects where the AI has to work for real people, not just look good in a demo.
Day to day I'm deep in Claude, LangGraph, n8n, and Vapi.
Check out my latest work
I've worked on a variety of projects, from simple automations to complex web applications and Agents. Here are a few of my favorites.

Whaply - Multi-Tenant WhatsApp AI SaaS
Production multi-tenant WhatsApp AI platform where each tenant gets a fully configurable AI agent handling inbound messages autonomously. Features LLM-powered intent classification, conversation routing, human-AI collision detection, and per-tenant agent configuration. Built with FastAPI, LangGraph for agent orchestration, Supabase/PostgreSQL for persistence, and Evolution API for WhatsApp integration — deployed on Railway with GitHub Actions CI/CD. Reduced per-conversation LLM cost by ~40% through Redis-backed intent caching and token budget enforcement.

LabLens MCP — Medical Lab Intelligence Server
Production MCP (Model Context Protocol) server that exposes four AI-powered tools for interpreting patient lab data — lab panel interpretation, critical value flagging, clinical summary generation, and differential diagnosis ranking. Supports SHARP Extension Specs for FHIR context propagation across multi-agent call chains, and integrates with Epic's FHIR R4 sandbox via OAuth 2.0 JWT assertion. Supports Anthropic Claude, OpenAI, Gemini, and Mistral as swappable LLM providers. Built for the Agents Assemble — Healthcare AI Endgame hackathon. Live at lablens.up.railway.app

Callahan Pipe & Drain - AI Voice Agent
Live AI voice agent for Callahan Pipe & Drain, a Chicago plumbing company. Embedded on the production site as a one-click call button — handles inbound service requests, emergency triage, and scheduling, then routes to human dispatch when needed. Built on Vapi with FastAPI backend integrations for calendar and CRM.

AddPost - AI-Powered Social Media Scheduler
Professional social media scheduling platform with AI-powered content generation for X (Twitter) and LinkedIn. Features AI caption generation, smart scheduling with timezone support, thread & carousel creation, PWA capabilities for offline use, secure authentication, and automated post execution with retry logic. Built with mobile-first approach and dark professional UI.

Klipit - AI Podcast Clipping Platform
AI platform that turns long-form podcasts into ready-to-post short-form clips. Transcribes episodes, identifies the most shareable moments (bold takes, contrarian opinions, emotional stories, how-tos), scores each clip 0–100 for engagement, generates hooks, and exports formatted for TikTok, Reels, Shorts, LinkedIn, and X. Ships with a REST API and webhooks for n8n / Make integration. Currently in active development.

Harmony Dental - AI Receptionist Voice Agent
HIPAA-compliant AI receptionist for Harmony Dental handling real patient workflows securely and autonomously. Features multi-step patient verification, secure appointment scheduling, prescription refills, lab results delivery, emergency detection, and automated reminders. Achieved 94% booking rate, 97% verification success, $2,400/month revenue recovery, 15 hours/week staff savings, 40% increase in new patient conversion, and 100% call answer rate with 24/7 availability. Live Demos: +1 (505) 356-8229 | +1 (507) 816-9529
ClearFlow Plumbing - Emergency Triage & Scheduling Voice Agent
AI receptionist for ClearFlow Plumbing handling emergency dispatch and service scheduling. Features real-time emergency classification (burst pipes, gas leaks, floods), automatic technician dispatch, dynamic pricing engine with upfront estimates, customer service history integration, and payment processing integration. Live Demo: +1 (702) 359-0236

Amara - Document to Audiobook
Personal project I built so I could listen to documents on the go. Upload a PDF, DOCX, TXT, or Markdown file and get back a clean MP3. Supports four TTS providers (Edge, Google, OpenAI, ElevenLabs), OCR fallback for scanned documents, async processing so uploads don't block the UI, and a history view with playback and download.
Work Experience
Skills
Get in Touch
Want to chat? Send me an email or shoot me a dm with a direct question on twitter and I'll respond whenever I can.