Make agent behavior legible
I design typed tools, context boundaries, retrieval flows, evaluations, and observability so teams can test and improve behavior instead of guessing.
About / Indra Zulfi Mushoddaq
Engineer, builder, and mentor.
I'm an AI Engineering Lead based in Surabaya, Indonesia. I build agent systems and full-stack AI products, usually where the problem is ambiguous, the context is dense, and the software still needs to be dependable in production.
My path has moved through visual systems, web products, engineering education, and AI. That range helps me connect low-level architecture to product behavior—and explain the decisions clearly enough for a team to keep moving.
Profile
Systems thinking with product range.
- Current
- AI Engineering Lead at AlphaX
- Based
- Surabaya, Indonesia; working globally
- Building
- Production AI systems, agent runtimes, and developer tools
- At scale
- 88K+ research reports and 45K+ analyses
- Education
- 600+ students across 24 Devscale cohort batches
My edge
I connect the whole system.
Own the path to production
I move between backend execution, data, interfaces, and deployment without losing the product outcome or the operational constraints.
Build systems that help teams move
I turn repeated engineering decisions into clear boundaries, reusable tooling, and practices that make the next feature easier to ship.
Working model
How I turn ambiguity into software.
- 01 / Frame
- Clarify the user outcome, system constraints, and evidence that will count as success.
- 02 / Bound
- Define contracts between models, tools, data, application code, and the people who operate them.
- 03 / Prove
- Build the smallest useful path with traces, evaluations, tests, and failure states from the start.
- 04 / Ship
- Connect the runtime to a usable product surface and make deployment repeatable.
- 05 / Improve
- Use production evidence to tighten behavior, reduce uncertainty, and simplify the next iteration.
Career arc
A cross-disciplinary path.
I started in visual and web work, grew into product engineering, founded an education platform, and now lead AI engineering. The medium changed; the habit of making complex systems understandable did not.
AlphaX
AI Engineering Lead
Japan · Remote · Aug 2024 - Present
Devscale Indonesia
Founder and Mentor
Indonesia · Hybrid · May 2023 - Present
Designstripe
Tech Generalist and Software Engineer
Canada · Remote · Aug 2020 - May 2024
SSH International
3D Generalist
Bahrain · On-site · Aug 2016 - Jul 2020
Pixellin Studio
Co-founder and Web Developer
Indonesia · On-site · Aug 2011 - Jul 2016
Toolbox
Tools I reach for.
I choose tools around the boundary and operating model, not around novelty. These are the systems I know well enough to make practical tradeoffs with.
Web development
- Frontend
- React, Next.js, TanStack, Svelte, HTMX, Tailwind CSS, Tailwind Variants, TypeScript, React Query, Three.js, Motion
- Backend
- Node.js, Bun, TypeScript, Express, Elysia, Hono, Python, Flask, FastAPI, Django
- Database
- Postgres, SQLite, MySQL, MongoDB, OracleDB
AI development
- LLM models
- Claude, OpenAI, Gemini, Llama, Mistral, Most leading LLM models
- Frameworks
- OpenAI SDK, OpenAI Agents, LangChain, LangGraph, LangSmith, PydanticAI, Agno, Vercel AI SDK, Building agents from scratch
- Vector DB
- pgvector, ChromaDB, Pinecone, Qdrant
- Observability and evals
- Langfuse, DeepEval, LangSmith
DevOps
- Tools
- Docker, Docker Compose, Kubernetes, Terraform, AWS (EC2, S3, RDS, Lambda), GCP, AWS, Azure, Alibaba Cloud, Tencent Cloud, CI/CD (GitHub Actions, Jenkins), Ansible, Bash scripting, Nginx
- Server / Linux
- Ubuntu, Debian, SSH, systemctl, Package management, Firewall management (ufw, iptables), Monitoring and logging (systemd-journald, syslog)
Best fit
Good problems to bring me.
- An AI product has outgrown the demo and needs a production architecture.
- Agent behavior is difficult to evaluate, observe, or improve with confidence.
- A small team needs one engineer who can connect runtime, backend, interface, and delivery.
- An engineering organization needs clearer tooling, technical direction, or practical AI education.