At Ardessa, I take backend features from prototype to production on AWS. I containerise Django and Celery services, define infrastructure in Terraform, and deploy through GitHub Actions.
I built a nightly asynchronous Python pipeline that aggregates Google Workspace and Microsoft Graph data through a unified collection layer. By consolidating per-section model calls into one, I cut per-item latency from 1m30s to 1m and reduced output errors.
At Ardessa, I also reduced model inference in a triage service by around 70% with a two-stage design. I designed a provider abstraction for three CRM APIs and built the Salesforce integration, including OAuth2 token refresh and a paginated SOQL client.
At SAGITAZ, I led the team integrating AI into the chatbot-building platform, enabling clients to deploy AI-driven chatbots. I also delivered CRM integrations and process automation for around 25 clients across the Zoho ecosystem, and spent nine months in Mexico City handling client-facing requirements analysis and API integration work.

