Senior Engineer at NinjaOne · Healthcare platform engineering at Hermina · Bandung, Indonesia
Backend and platform engineer working across Ruby on Rails, Go, and Python services in production healthcare and IT-operations environments. Alongside product work, I build the retrieval and memory infrastructure that makes AI-assisted development reliable at multi-repository scale.
Platform engineering. Service backends on Ruby on Rails 8 and Go with the Kratos framework, paired with React, Next.js, and React Native clients. PostgreSQL, Kubernetes, and AWS underneath.
Tooling for AI-assisted engineering. Context and retrieval infrastructure for coding agents: cross-agent session memory, code knowledge graphs, and MCP servers. The premise is that agent quality is bounded by the context you can supply, which makes context an engineering concern rather than a prompting one.
agent-memory — A cross-agent memory layer. It ingests session history from six coding agents (Claude Code, Codex, OpenCode, Kimi CLI, DeepSeek, Copilot), normalizes it, resolves each session to its originating repository, and indexes the result in PostgreSQL with pgvector alongside a human-readable Obsidian vault. Hybrid retrieval is exposed through a CLI and an MCP server so every agent operates from shared memory. Over 1,700 sessions indexed to date. Python, Typer, SQLAlchemy, Alembic, pgvector.
Code knowledge graph — An incrementally updated, Tree-sitter-parsed graph of the codebase, used for impact-radius analysis, caller and callee tracing, and token-efficient review context.
codex-lb — Load balancer and proxy for pooled ChatGPT and Codex accounts, providing usage tracking, API key management, an operations dashboard, and OpenCode-compatible endpoints. Python, FastAPI, SQLAlchemy, OAuth.
| Area | Technologies |
|---|---|
| Backend | Ruby on Rails, Go (Kratos), Python (FastAPI), Node.js |
| Frontend | React, Next.js, Vite, Ant Design, Tailwind CSS |
| Mobile | React Native, Expo |
| Data | PostgreSQL, pgvector, Redis, Elasticsearch |
| Infrastructure | Docker, Kubernetes, AWS, GitHub Actions |
| AI systems | MCP servers, RAG and hybrid retrieval, multi-agent orchestration |
