"I build AI agents and automation systems that eliminate manual work — from idea to live deployment"
I am an AI Automation Engineer based in New Hampshire, USA. I design and deploy multi-agent AI systems and automation workflows that eliminate manual work and connect AI to real business tools.
My work spans building agentic AI assistants, voice bots, document automation tools, and cloud-based AI applications using n8n, OpenAI, AWS Bedrock, and MCP. I focus on shipping live, working products — not just prototypes.
I hold AWS Solutions Architect and AWS AI Practitioner certifications and bring 12+ years of professional experience across media, legal, and technology industries.
Problem: Professionals waste hours every day switching between Gmail, Google Calendar, and meeting tools. Every task lives in a different app and there is no single place to manage it all.
Solution: Maestro is a multi-agent AI assistant that handles email, calendar, and meeting tasks through one conversational chat interface. A Master Agent receives the user's message, identifies the intent, and routes it to one of three specialized sub-agents — each an expert in its own domain — which take action automatically.
Key Outcomes:
• 3 specialized AI agents deployed across Gmail, Google Calendar, and Fireflies.ai
• Integrated 12 live API actions — read, send, draft, reply, delete emails, get, create, update, delete calendar events, and retrieve meeting transcripts
• Reduced workflow timeout errors by 90% through API limit optimization
• Collapsed 3 separate productivity tools into one conversational interface
• Live and accessible via public URL — no login or app install required
Problem: Women's health clinics lose patient bookings after hours when calls go unanswered. Patients move to competitors and staff return to missed calls with no structured data.
Solution: An AI voice agent named Maya answers every inbound patient call 24/7, collects booking details, checks real-time availability, schedules appointments into Google Calendar, and logs all post-call data to Google Sheets — automatically.
Key Outcomes:
• 100% of patient calls answered
• 24/7 availability — no staff needed
• 0 minutes wait time for booking
• 100% of post-call data captured automatically
One photo. Instant safety report. Zero manual inspection.
Problem: Workplace safety inspections are slow, inconsistent, and expensive. Manual PPE checks require trained staff on-site and still miss violations that put workers at risk.
Solution: VisionCheck AI analyzes any workplace photo and generates a full PPE compliance report in under 15 seconds — automatically detecting violations, rating risk level, and recommending corrective actions.
Key Outcomes:
• PPE compliance report generated in under 15 seconds
• Risk level automatically rated as Low, Medium, or High
• Corrective actions recommended without human review
• Live and accessible at visioncheck.buildsmartai.app
Your contract reviewed in 30 seconds. No lawyer needed.
Problem: Contract review costs $300–$600 per hour and most small businesses, startups, and freelancers simply cannot afford professional legal review before signing.
Solution: RedlineAI reads any legal contract, identifies every risky clause, explains each one in plain English, assigns an overall risk level of Low, Medium, or High, and delivers a professional risk report directly to the user's inbox in under 30 seconds — fully automated, no human involved.
Key Outcomes:
• Contract analyzed and risk report delivered in under 30 seconds
• Risky clauses identified and explained in plain English
• Two modes: automated pipeline and Claude MCP agent mode
• Live and accessible at redlineai.buildsmartai.app
Your destination. Your style. Your plan — in 2 minutes.
Problem: Travel planning is slow, fragmented, and overwhelming. Travelers spend hours switching between blogs, maps, YouTube videos, and spreadsheets just to build a basic itinerary — often ending up with generic plans.
Solution: WANDR is an AI-powered travel planning web app that transforms a destination idea into a complete personalized travel plan in under 2 minutes. It generates day-by-day itineraries, hidden gems, packing lists, food recommendations, and safety tips — all through one guided experience.
Key Outcomes:
• Complete travel itinerary generated in under 2 minutes
• Day-by-day plan with morning, midday and evening activities
• PDF export, email delivery, and public shareable trip links
• Built and shipped as a fully responsive web app using Lovable (vibe coding)
AI that drafts. Humans that approve. Customers that get faster, better answers.
Problem: D2C skincare support teams spend 8-12 minutes manually drafting policy-based email replies — causing slow response times, inconsistent answers, and agent burnout.
Solution: Built an AI-powered customer support workflow using n8n, Gmail, OpenAI, Supabase Vector Store, and Slack to automate email triage, policy retrieval, and response drafting — with human review before anything reaches the customer.
Key Outcomes:
• Reduced draft time from 8-12 minutes to 2-4 minutes
• Replies grounded in company policy using RAG — not hallucinated
• 100% human review maintained before customer contact
• Support capacity increased without adding headcount
• Real Supabase Vector Store in production — project: novaglow-ai-support
I am open to full-time AI Automation Engineer roles, freelance AI automation projects, and collaboration on agentic AI systems. Reach out directly or send a message using the form.
Since you are on a desktop, choose one of these options to call Maya:
Scan with your phone's camera to dial instantly:
Dial from your phone or softphone app: