Freelance AI Programmer: Engineering Autonomous LLM Agents, Hybrid RAG & MCP Servers

Building an AI prototype in a notebook takes an afternoon; engineering a reliable, secure, cost-controlled production AI application requires a seasoned Freelance AI Programmer and AI Application Engineer. From deterministic tool calling and sandboxed code execution to hybrid vector-graph memory, production AI systems succeed or fail on their systems engineering.

I am Aldian Fazrihady, creator of Journal: Private Markdown Notes, Transcribe Copilot, and Interview Copilot, and a Top-Rated Freelance AI Programmer on Upwork (aldianf).

Production AI Engineering Services

• Autonomous AI Agents & Sandboxed Execution: Ephemeral Docker sandboxes for AI coding and operations agents (Claude Code, Cline, NanoClaw), bridging ChatOps (Microsoft Teams, Slack) with secure container runtimes.
• Model Context Protocol (MCP) Architecture: Designing custom MCP tools, stdio-to-SSE bridges, and multi-tenant authorization layers so LLMs can safely query internal ERPs, CRMs, FHIR servers, and SQL databases.
• Dual-Memory RAG & Knowledge Graphs: Combining short-term conversation context with asynchronous NATS fact extraction, PostgreSQL pgvector embeddings, and FalkorDB graph memory.
• Multimodal Vision, Voice & Document AI: Healthcare insurance card extraction with AWS Bedrock Vision, automated tax invoice OCR with Google Cloud Document AI, and real-time audio pipelines.

Featured AI Engineering Articles

• The Dual-Memory Architecture: Orchestrating Short-Term Agent Context and Long-Term Knowledge Graphs
• Standard I/O to SSE: Bridging Ephemeral Docker Sandboxes to Host MCP Servers
• Securing AI Coding Agents: Building Ephemeral Docker Sandboxes for Claude Code and Cline
• Vision AI in Healthcare: Building an Insurance Card Extractor with AWS Bedrock and Medplum Binaries