I build AI automation and the full stack it runs on.

Autonomous revenue intelligence

A multi-tenant revenue intelligence platform that works for a sales team around the clock. It finds prospects that match the roles, industries, and regions a business chooses, writes to them in their own language from the business's own Google Workspace or Microsoft 365 mailboxes, and turns replies into qualified pipeline, all within the rules the business sets.

I started this project from scratch and built its core backend, which I still own end to end, and I work on its Next.js web app and platform-admin panel.

• Organizations with database-backed roles and Postgres row-level security, enforced through a non-superuser app role
• Prospecting from target criteria through discovery, channel identification, qualification, escalation, and scheduled outreach, run by Celery workers
• Multilingual outreach and nurture sequences through Amazon SES, with engagement tracking, opt-out suppression, and reply detection that stops follow-ups when a prospect answers
• Autonomous conversation handling on a LangChain LLM layer that switches between Anthropic and OpenAI, with an audit of every decision
• Deployment to AWS ECS through GitHub Actions: a test gate, migrations in a one-off Fargate task, then rolling services forward

• FastAPI
• PostgreSQL
• SQLAlchemy
• Celery
• Redis
• LangChain
• Amazon SES
• AWS ECS
• GitHub Actions
• Next.js
• React

Enterprise change management

Backend for an enterprise change-management platform. It tracks stakeholders, initiatives, and department-level readiness and impact across an organization.

Built on FastAPI and PostgreSQL, with SQLAlchemy models and Alembic migrations.

• Models cultural resistance and sentiment across an organization, and scores training adoption
• AI-assisted endpoints that generate scenario recommendations
• A reporting layer that exports findings to PDF, Excel, and PowerPoint

• FastAPI
• PostgreSQL
• SQLAlchemy
• Alembic
• Docker

Beauty and business management

Backend for a beauty and business management platform used through a mobile app. It handles bookings, catalog, and scheduling, plus payments end to end.

I designed and built the Classes module (regular classes and workshops) and the audit logging module, and fixed timezone handling issues across the project.

• A full Finix payment integration: card tokenization, subscriptions, webhook-driven reconciliation, and dispute handling
• Postgres through Supabase, with Alembic-managed migrations
• Bookings, catalog, and scheduling served to the mobile app

• FastAPI
• SQLModel
• PostgreSQL
• Finix
• Docker