AI integration engineer. 20+ years building production systems that don't fall over. Cloud architecture, RAG pipelines, autonomous agents, MCP servers, and HIPAA-compliant healthcare AI.
I started writing code in the late 90s and never stopped. My first job was building Java enterprise systems for telecom companies in Jakarta. Somewhere along the way I picked up a Master's in Electronic Engineering from ITB, got Oracle-certified in Java, and spent a few years doing distributed systems work before cloud was even a word people used seriously.
These days I split my time between two things: cloud architecture and AI integration. They overlap more than you'd think.
On the cloud side, I've been building on AWS, GCP, and Azure for over a decade. I'm Google Cloud certified as both a Professional Cloud Architect and a Professional Data Engineer, which I earned because I was already doing the work and figured I might as well have the paper to prove it. Most of my projects involve Terraform, containerization, serverless, and the unglamorous operational work that keeps systems running when traffic spikes at 3am.
On the AI side, I build the integration layer between LLMs and real business systems. RAG pipelines that pull from actual company documents instead of generic training data. Autonomous agents that follow clinical workflows, check drug interactions against FHIR resources, and know when to hand off to a human. MCP servers that give language models structured reach into databases and APIs. I work with OpenAI, Claude, Gemini, and self-hosted models through vLLM, depending on what the client needs and what their budget looks like.
The healthcare work is where I've spent the most time recently. HIPAA-compliant architecture for medical SaaS platforms, clinical decision support systems, drug safety checking. Real systems processing thousands of requests a day, not demos. Before that I built IoT backends handling over a thousand writes per second and backend systems for a few startups that either got acquired or quietly disappeared, which is how that industry works.
My tech stack: Python (FastAPI) for most API work, Go (Gin/GORM) when performance matters, TypeScript and React for frontend, PostgreSQL and BigQuery for data. I use Terraform for everything infra-related and Docker for anything that needs to run consistently across environments.
I've been freelancing for over 15 years. The thing I've learned is that communication matters as much as code. Clients don't care about your elegant architecture if you can't explain what you're building and when it'll be done. So I keep things simple, ship often, and say "I don't know" when I don't know, then figure it out.
Based in West Java, Indonesia. I work with clients across timezones and I'm flexible about it.
Work Terms
Payment: All work goes through Guru's SafePay (escrow). For ongoing relationships we can switch to weekly invoicing once trust is established.
Communication: I'm in UTC+7 (West Java). For clients in Europe or North America, some messages will be overnight, but I reply within 12 hours max. Calls by appointment. I'm fine with async if that's your thing. The key is clarity: if there's uncertainty about what to build, let's resolve it before I write code.
Availability: Monday through Friday, 20-30 hours most weeks depending on project stage. Urgent deadlines happen and I get that, just don't make it a habit.
Scope: I prefer starting with clear scope: features, constraints, acceptance criteria. If you don't have one, I can help write it during onboarding. Changes mid-project are normal but need a conversation about timeline and cost impact. Scope creep burns out good engineers.
Revisions: Code review feedback is expected and welcome. If I misunderstood a requirement, I'll fix it. But "this isn't what I visualized" three weeks after delivery is new work, not a revision. We handle that through change orders.
Confidentiality: Healthcare work means HIPAA. I've built HIPAA-compliant systems and know what's required. Comfortable signing NDAs and working through your security review. For healthcare projects, I'll flag any architectural decisions that could affect compliance before implementation.
Ownership: You own what you pay for. Delivered code, docs, and assets are yours. I don't reuse client code across projects. The only exception is code explicitly identified as reusable from the start, and even then we agree on licensing upfront.
Project closeout: When work is done, I provide a handover document with architecture overview, deployment instructions, and known issues. A short post-launch support window to handle bugs can be built into the contract.