Track 01 · AI Automation
AI Automation
Repetitive work, off the team’s plate. AI agents, no-code flows, and custom integrations that pay for themselves — measured in hours, not hype.
faster product upload vs the manual workflow
automation flows in production across clients
reclaimed per month by clients on average
Featured case
From 90 minutes to 15 minutes
AI agents that bulk-upload products, write SEO copy, and ship to marketplaces — what used to be a 90-minute manual task now finishes in 15.
See the full case study →What I build
Workflow Audit
I sit with your team and map every repetitive step — from inbox triage to onboarding to fulfillment. The audit ends with a prioritized backlog of automations ranked by ROI.
AI Agents & Chatbots
Conversational agents that answer customer questions, qualify leads, and route work to the right human. Built on OpenAI, Anthropic, or open-source models via streaming APIs.
n8n & Make Automations
No-code and low-code flows that connect your CRM, email, calendar, sheets, and SaaS tools. Trigger to action, end to end, with logging and retry handling.
E-commerce Automation
Bulk product uploads, AI-generated descriptions and images from phone snapshots, multi-platform publishing. The 90→15 minute case study you saw in the portfolio — engineered for real stores.
Lead Generation & Nurture
AI-driven scraping, enrichment, and personalized cold outreach that reads like a 1:1 message. Warm leads land in your CRM, ready for your sales team to close.
Custom Integrations
Webhooks, REST and GraphQL APIs, OAuth flows, background jobs. When the SaaS you need doesn’t exist, I build the bridge that does the job.
How I work
Five steps, every project. The same rhythm whether the deliverable is a saved hour or a shipped product.
Discover the workflow
Every project starts with the workflow, not the tool. I dig into how your team actually works today, where time is leaking, and which automations pay for themselves in weeks — not quarters.
Ship in slices
Once the backlog is locked, I ship in small, observable pieces. Every flow has a dashboard, every error has an alert, every retry has a story.
Prompts as code
I treat prompts like code: version-controlled, evaluated against golden datasets, and rolled back when they drift. No black boxes on my watch.
Data stays put
Your data stays where it belongs. Self-hosted models when privacy or cost demands it, third-party APIs when speed to ship matters more.
Iterate from real traffic
After launch, I monitor what real traffic teaches and iterate. The first version is the start of the conversation, not the end of the project.
My stack
- OpenAI / Anthropic / Ollama / vLLM
- n8n / Make / Zapier
- Telegram / Slack / WhatsApp Business APIs
- Airtable / Notion / Google Sheets
- HubSpot / Pipedrive / Close / Brevo / Mailchimp
- WooCommerce / Shopify / Stripe webhooks
- PostgreSQL + Prisma / Drizzle
- Docker / self-hosted VPS
Unleash Potential Together
Ready to stop doing manually what could be automated in seconds? Let’s map the first workflow.
Book a free strategy call