Hasaif Akram

Designed and deployed production-grade n8n automation workflows integrating Google Gemini, PostgreSQL, Pinecone, Supabase, and Google Workspace APIs to solve real operational problems across customer support, finance, and content operations — not proof-of-concept demos, but systems built with the validation, error-handling, and safety gates production use requires. Built a PostgreSQL-backed conversational AI admin agent spanning 7 business domains (branches, menu, inventory, orders, returns, sales), with strict enum validation, UUID resolution before any write operation, confirmation gates on destructive actions, and rolling conversation summarization to manage long-term memory without exceeding context limits. Engineered a bilingual (English/Arabic) RAG chatbot with automatic language detection, routing each query to a language-specific agent and an independently indexed Pinecone knowledge base — using Ollama embeddings and language-aware system prompts to keep responses grounded and policy-compliant rather than hallucinated. Built a Gmail-triggered customer support agent combining AI-based intent classification with Retrieval-Augmented Generation to auto-draft context-aware, policy-compliant replies, cutting manual triage effort for the support team. Implemented a social content automation pipeline: RSS feed monitoring, duplicate detection, AI-generated LinkedIn posts, automated publishing via the LinkedIn API, and status tracking through Google Sheets — a full content-ops loop with no manual intervention required. Core competencies demonstrated: agentic workflow design, RAG architecture, multi-domain conversational AI, API integration, prompt engineering, database-safe automation design, and end-to-end business process automation.

Bilytica