Luko AI
FlagshipAI chatbot for students asking about informatics topics. Integrates LLM models with a Retrieval-Augmented Generation (RAG) system for document-grounded answers.
LLM + RAG Document Grounding + Async Inference Queue
Backend & AI Systems Engineer · Go · Python · RAG
Building systems that scale — from high-concurrency Go backends and async message brokers to document-grounded Retrieval-Augmented Generation.
Production-oriented projects across distributed Go backends, RAG pipelines, Proxmox cloud IaaS, and automated MLOps.
AI chatbot for students asking about informatics topics. Integrates LLM models with a Retrieval-Augmented Generation (RAG) system for document-grounded answers.
LLM + RAG Document Grounding + Async Inference Queue
Official Online Judge at Universitas Tanjungpura. Automated code assessment backend built with Go, PostgreSQL, RabbitMQ, and sandboxed execution workers.
Worker Pool + Isolated Sandbox + AMQP Task Routing
IaaS/PaaS cloud provisioning platform powered by Go, Proxmox VE APIs, Valkey caching, OAuth2/JWT, and a local payment gateway.
Proxmox API + JWT Auth + Local Payment Gateway
MLflow CI/CD pipeline: automated retraining and deployment of a RandomForest model (breast cancer classification) with GitHub Actions tracking accuracy, F1, precision, and recall.
Automated Retrain + Model Registry + Metrics Validation
How I design reliable APIs, ground language models in verified documents, decouple heavy compute, and operate infrastructure.
Informatics Engineering student at Tanjungpura University (UNTAN), focused on backend engineering and AI engineering.
Building APIs & backend systems with Go, PostgreSQL, and RabbitMQ — plus RAG (Retrieval-Augmented Generation) systems for AI chatbots. I believe great systems are observable, resilient, and measurable.
Designing high-throughput Go services, tuned PostgreSQL connection pools, and low-latency REST APIs built for predictable performance under load.
Integrating language models with semantic vector retrieval pipelines so chatbots answer strictly from verified domain documents with minimal hallucination.
Offloading compute-heavy workloads — automated code judging and LLM inference — into isolated worker pools via RabbitMQ and Kafka.
Automating Proxmox VE hypervisor provisioning alongside reproducible MLflow training, registry validation, and GitHub Actions deployment workflows.
Trace end-to-end distributed request flows across Go API gateways, RabbitMQ brokers, and isolated workers — or benchmark in-memory cache latency.
Select a pipeline above and click "Run Trace" or inspect any node.
Deep dives on building physical homelab infrastructure, cross-border supply chains, and software engineering project lifecycles.
A 3-month deep dive into architecting a physical high-compute server: China DAP supply chain sourcing, electrical grid certification hurdles with PLN, and advanced server software provisioning.
Why software projects fall apart not because the code was too hard, but because the process was chaotic. A grounded look at the 5 Project Life Cycle phases, Waterfall vs Scrum vs XP trade-offs, and pragmatic hybrid setups.
Architecting a Retrieval-Augmented Generation knowledge engine for UNTAN Informatics students with vector embeddings, chunk retrieval, and async queues.
Building high-concurrency REST APIs for automated code grading across thousands of submissions with RabbitMQ queues and PostgreSQL.
Developing a VM provisioning platform powered by Proxmox VE APIs, Valkey caching, OAuth2/JWT auth, and local payment gateway webhooks.
Automating MLflow retraining, model registry validation, and deployment workflows with GitHub Actions.
Focusing on distributed systems architecture, algorithms & data structures, POSIX operating systems, and applied AI engineering.
Interested in collaborating on backend systems, AI/RAG pipelines, or infrastructure? Send a note.
Interested in collaborating on backend systems, AI/RAG pipelines, or infrastructure? Send a note.
Open for backend engineering, distributed systems, and AI/RAG engineering roles & collaborations.