Looks like you have JavaScript disabled. For the full Twine experience, you will need to re-enable it.

AI Automation Engineer architecting production-grade systems around a 20-year-old POS, syncing bidirectionally with PostgreSQL and delivering staff and customer-facing AI experiences (WhatsApp agent). Builds and operates unattended 24/7 automation and AI agents with n8n, RAG pipelines, vector search, and evaluation/benchmarking, along with full-stack products (mobile + React Native) and hybrid integrations between legacy POS and cloud services.…AI Automation Engineer architecting production-grade systems around a 20-year-old POS, syncing bidirectionally with PostgreSQL and delivering staff and customer-facing AI experiences (WhatsApp agent). Builds and operates unattended 24/7 automation and AI agents with n8n, RAG pipelines, vector search, and evaluation/benchmarking, along with full-stack products (mobile + React Native) and hybrid integrations between legacy POS and cloud services.

GusDev

AI Engineer, Full Stack Developer, Web Developer, +3





AI Automation Engineer architecting production-grade systems around a 20-year-old POS, syncing bidirectionally with PostgreSQL and delivering staff and customer-facing AI experiences (WhatsApp agent). Builds and operates unattended 24/7 automation and AI agents with n8n, RAG pipelines, vector search, and evaluation/benchmarking, along with full-stack products (mobile + React Native) and hybrid integrations between legacy POS and cloud services.…AI Automation Engineer architecting production-grade systems around a 20-year-old POS, syncing bidirectionally with PostgreSQL and delivering staff and customer-facing AI experiences (WhatsApp agent). Builds and operates unattended 24/7 automation and AI agents with n8n, RAG pipelines, vector search, and evaluation/benchmarking, along with full-stack products (mobile + React Native) and hybrid integrations between legacy POS and cloud services.

Available to hire

AI Automation Engineer architecting production-grade systems around a 20-year-old POS, syncing bidirectionally with PostgreSQL and delivering staff and customer-facing AI experiences (WhatsApp agent). Builds and operates unattended 24/7 automation and AI agents with n8n, RAG pipelines, vector search, and evaluation/benchmarking, along with full-stack products (mobile + React Native) and hybrid integrations between legacy POS and cloud services.

Skills

Experience Level

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Intermediate

Intermediate

Beginner

Beginner

Beginner

Language

Spanish; Castilian

Fluent

English

Advanced

Work Experience

Inventory & Sales Analytics Developer (AI anomaly detection) at El Serrucho GoGo

July 1, 2026 - Present

Built daily sales analytics with full invoice history and live inventory workflows for staff. Added quote building with PDF export and enabled purchases to queue into the POS through a write-back bridge mechanism. Shipped Android and PWA from an Expo/React Native/TypeScript codebase. Designed a three-tier role model with approval gates and implemented anomaly detection using Gemini, supporting inventory/sales monitoring workflows while ensuring requests are queued rather than directly rewritten to the POS.

AI Agent & Inventory/Sales Analytics Developer at El Serrucho Go

June 1, 2026 - Present

Built daily sales analytics including full invoice history and live inventory for staff. Added quotation building with PDF export and queued purchases to be written back into the POS via bridge listeners. Implemented a three-tier role model with approval gates and Gemini-based anomaly detection. Shipped an Android app and PWA using an Expo/React Native/TypeScript codebase. Designed it to avoid direct POS writes by enqueuing requests and replaying through bridge listeners.

Perucho - Autonomous WhatsApp AI Agent (AI Systems in Production) at Perucho

June 1, 2026 - Present

Built and operated an autonomous WhatsApp AI agent on n8n and Docker (WAHA, OpenAI, Groq Whisper). Designed a 5-layer retrieval pipeline over ~7,650 SKUs with common query resolution around ~900ms. Served 191 customers since June with high containment (only 37 conversations required a human). Built a 320-case evaluation benchmark for agent failures, improving exact recall from 72.2% to 76.9%. Increased vector signal-to-noise by embedding enriched text rather than raw catalog data. Delivered hybrid on-cloud bidirectional bridging between legacy POS (Delphi/BISAM POS) and Supabase, including BISAM data parsing and robust write-back through the POS UI.

WhatsApp AI Agent Builder / AI Systems in Production (n8n + Docker) at Perúcho - Atónonomous WhatsApp AI Agent (GitHub: /Gus2708/whatsapp-agent)

June 1, 2026 - Present

Built and operated an autonomous WhatsApp AI agent using n8n and Docker with LLM providers (OpenAI and Groq Whisper). Implemented a 5-layer retrieval pipeline over ~7,650 SKUs with common query resolution in ~900ms. Evaluated the agent with a ~320-case benchmark dataset; when the agent failed, exact recall improved from ~72.2% to ~76.9%. Enhanced vector search signal-to-noise using enriched text embeddings instead of raw catalog data.

AI Systems in Production - Hybrid on-cloud bidirectional bridge at Hybrid Bridge (Legacy POS <-> Supabase)

May 1, 2026 - Present

Built a hybrid bridge connecting a closed Delphi/D-BISAM POS to Supabase, running unattended 24/7 since May. Parsed BISAM data tables in Python as read-only streams and wrote back through the POS UI using a Windows SendInput workaround after synthetic input reliability issues. Implemented sync-on-file-change detection using MD5 checksums and isolated multi-process queue listeners/supervisor logic behind mutexes for safe production synchronization.

Full Stack Mobile Developer at Caoba Café (client project)

March 1, 2026 - Present

Delivered a data-first POS for a café client, scoped and shipped directly with the owner. Implemented offline-first inventory with real-time Supabase sync supporting multi-tenant RLS and PDF invoicing. Built and shipped the client’s mobile app with React Native/TypeScript and corresponding tooling/infrastructure. Focused on direct inventory and invoicing integration to reduce friction in day-to-day cafe operations.

Full Stack Mobile Developer at Caoba Café

March 1, 2026 - Present

Delivered a first-to-market POS for a café client and supported shipment directly with the owner. Implemented offline-first inventory, real-time Supabase sync for a multi-tenant setup using RLS and edge functions, and PDF invoicing workflows. Developed for Android and provided PWA installation support, using a React Native / Expo / TypeScript stack and maintaining the codebase via GitHub.

Operations Manager & In-House Developer at Ferretería El Serrucho M&F.C.A.

May 1, 2023 - Present

Modernized a 20-year-old on-premise POS into a cloud platform. Automated invoice data entry by removing manual work from the POS UI (cutting entry time from up to 2 hours to ~30 minutes). Built a Python bidirectional bridge syncing ~7,650 SKUs with PostgreSQL and used the database for a staff app plus a customer-facing AI agent. Managed 70+ daily transactions and a 7-person operations team; later automated the operational workflow. Developed full-stack capabilities including a staff-facing mobile application and integrations for inventory and sales workflows, using RLS/Edge Functions and syncing via Supabase. Implemented reliable scheduling (file-change watch + MD5 checks) and ensured production stability with isolated queue listeners/supervisors behind mutual exclusion.

AI Automation Engineer - n8n, AI Agents & RAG in Production at Ferretería El Serrucho M&FC.A

May 1, 2023 - Present

Modernized a 20-year-old on-premise POS into a cloud platform powering day-to-day operations. Developed a Python bidirectional bridge syncing data with PostgreSQL and built a React Native staff app plus a WhatsApp AI agent for customers. Kept services running unattended 24/7, moved data entry from the POS UI, and automated invoicing and inventory-related workflows. Designed and operated n8n AI agent workflows including tool calling, workflow monitoring, webhooks, and RAG pipelines with vector search (pgvector). Implemented multi-step retrieval and evaluation/benchmarking to improve exact recall rates. Built REST API integrations, queue/listener patterns, and change-driven syncing (MD5-based watchdog) to sync on file changes rather than schedules. Used Docker and CI/CD practices while focusing on production monitoring and reliability.