Software Engineer - Backend, Distributed Systems and AI/LLM
Wigtono
I build distributed systems and the retrieval systems behind production AI/LLM applications.
I build backend and distributed systems, and production retrieval infrastructure: ingestion, hybrid retrieval, reranking, context selection, evaluation and cost control, in Go, Rust, Python and TypeScript.
- Role
- Software Engineer - Backend, Distributed Systems and AI/LLM
- Experience
- Nearly three years across backend and AI/LLM systems
- Open to
- Backend, distributed systems and AI/LLM engineering roles, including retrieval and RAG infrastructure
- Work
- On-site in Indonesia · Remote, timezone-flexible
- wigtono41@gmail.com
- Contact
- Profiles and how to reach me
At a glance
- Focus
- Backend, distributed systems and AI/LLM systems
- Platform scale
- 48+ merchants and 20K+ outlets on one loyalty CMS
- Open to
- Software engineering roles - on-site in Indonesia, or remote
Public scope
Professional work
Selected client work at publishable scope: the product, the engineering problem, and the decisions behind it.
Professional workPinned
GANI.AI
Retrieval infrastructure behind a production legal AI platform for legal, tax, and compliance work.
- Retrieval infrastructure for legal, tax, and compliance workflows
- Embedding, retrieval, reranking, and context selection
Professional workPinned
Janji Jiwa Multi-Brand Apps
One app suite for every Jiwa Group brand: customer ordering, outlet operations, and an offline-capable rider app for a network of 900+ outlets.
- One customer app covering every brand under Jiwa Group
- Rider app that sells offline and syncs when the connection returns
Professional workPinned
ZU Point Loyalty System CMS
A loyalty CMS for a rewards network of 48+ merchants and 20K+ outlets: voucher lifecycle, promo code engine, audit logs, and reporting.
- 48+ merchants and 20K+ outlets on one voucher platform
- Redemption that has to be exactly once under concurrency
Implementation detail
Open-source work
Systems and libraries where the repository, architecture, and tradeoffs can be inspected directly.
Open sourcePinned
Pi-Herdr
A durable multi-agent orchestration extension for Pi and Herdr, with a SQLite control plane, durable mailboxes, workflow DAGs, and structured completion.
- SQLite control plane for agent identity, assignment, and state
- Durable mailboxes between parent and worker agents
Open sourcePinned
Sadewa
A type-safe Go library for integrating multiple trading exchanges, covering order models, streaming, rate limiting, circuit breaking, and reconnection.
- One interface across Binance and Bybit
- Streaming with reconnection as part of the data model
Focus areas
Designing and evolving services that coordinate across boundaries, with attention to failure modes, consistency, and operability.
See related work ->Production RAG infrastructure - hybrid retrieval, reranking, context selection, retrieval evaluation, and the cost of running it.
See related work ->Service-oriented backends with clear boundaries, maintainable interfaces, and room to evolve.
See related work ->
Engineering toolkit
More about me ->Languages
Languages I use to build services, libraries, and tooling.
- Rust
- Go
- Python
- TypeScript
Systems and infrastructure
The engineering space I work in and want to keep deepening.
- Microservices
- Distributed systems
- Docker
- Kubernetes
AI / LLM systems
RAG, knowledge systems, and service architecture for practical AI products.
- RAG and hybrid retrieval
- Reranking and retrieval evaluation
- AI microservices
- Agent orchestration
- Knowledge systems
Selected writing
All writing ->Notes on engineering, systems, and the work behind the projects.
Designing a Production RAG Retrieval Pipeline
A practical architecture for hybrid retrieval, reranking, access-aware context assembly, evaluation, and cost control in production RAG systems.
Discuss an opportunity
Email is the most reliable way to reach me: wigtono41@gmail.com. Profiles and other channels are on the contact page.