Juan Diego Gómez

Enterprise hospitality integrations across the Oracle Opera Cloud ecosystem (OHIP) and Salesforce CRM, owning the architecture that connects PMS and CRM platforms.

• Designed and owned the integration architecture between Oracle Opera Cloud PMS (OHIP) and Salesforce CRM: real-time sync of guest profiles, reservations, loyalty data and customer lifecycle information.
• Built an MCP tool layer exposing hotel operational systems to LLM agents via structured APIs, enabling agent-driven booking, guest and service operations.
• Integrated ElevenLabs voice agents into hotel operations for real-time conversational staff-facing and guest-facing interfaces.

Automation and payment systems for international hotel chains including Marriott, Atlantis and Pestana, focused on reservations and guest profile workflows.

• Designed Python services for credit-card check-in and deposit capture between FreedomPay and Opera Cloud PMS.
• Built idempotent REST APIs and event-driven jobs (Docker + Kubernetes on OCI) for real-time reconciliation.
• Created invoice and tax automation pipelines for hospitality groups including Homa Apartments and Pestana Hotels.

An AI-driven marketplace automation platform serving 10,000+ users — pricing recommendations, listing generation and negotiation messaging.

• Shipped LLM-powered systems combining OpenAI APIs with rule-based guardrails for reliability and controllability in production.
• Engineered automation pipelines for listing generation, pricing optimization and messaging workflows.
• Grew to 10,000+ users before acquisition.

Supported development and deployment of a speech-to-text and meeting summarisation product with Zoom and Teams integrations.

• Ran technical deployments and integration work for Zoom/Teams customers.
• Bridged customer requirements and the engineering roadmap for STT pipelines.

Marketplace automation for 10,000+ resellers: listings from photos, pricing off live comparables and negotiation messaging, sold as a €20/month community on Whop.

• Python around OpenAI APIs with deterministic price and inventory guardrails.
• Apparel data: sizing and colourway normalisation, condition grading, counterfeit signals.
• 5.0 ★ over 426 public reviews; acquired in 2026.

A Silicon Valley AI company whose product listens to sales calls on Zoom, Google Meet and Microsoft Teams and learns from each one — reading client attitude, surfacing company intelligence and feeding it back to the team.

• Speech-to-text and meeting summaries deployed into Zoom and Teams.
• Sentiment and attitude signals per call, turned into coaching for the sales floor.
• Sold and scoped the integrations customers actually ran.

My degree project: a classifier that separates automated Instagram accounts from real people using public profile, posting and engagement signals.

• Feature set built from posting cadence, follower/following ratios, caption reuse and engagement shape.
• Labelled dataset, train/test split and a supervised model scored on precision and recall, not raw accuracy.
• The interesting part is the cost asymmetry: flagging a real person is worse than missing a bot.