M. Daffa Raygama ✦
Building production LLM systems and real-time AI pipelines at PT Sigma Cipta Utama. Telkom University CS grad, 3.96 GPA. Passionate about open-source AI tooling and multi-agent architectures.
Featured Project
the one i'm most proud of
Solo Project · 5 months
AURA — Affective Understanding & Responsive Avatar
A delusion made real.
A real-time AI companion with emotion-aware responses, Live2D avatar, and full WebRTC voice pipeline. Deepgram STT → DeepSeek V3.2 LLM → Qwen TTS → LiveKit.
Case Studies
Rizzy Bytes
A production RAG platform serving 1,000+ university students. Built with 7 microservices, async message queues, and a multi-modal knowledge base supporting 25+ KB docs. Led architecture and full-stack development solo.
Landing page
Chat interface
Knowledge base
User management
Monitoring2025 · PM + Full-stack · Restaurant Platform
Sunkatsu App
Digital ordering platform for Sunkatsu, a Japanese restaurant in Telkom University's campus cafeteria. Spans web, mobile, and a Qwen-2.5 AI chatbot. Led a team of 6 as PM and handled frontend architecture delivered with 11 language localizations and full ordering flow across all platforms.
Home
AI chat
Orders
Mobile menu
Mobile cart2026 · AI Engineer · Database Agent & Eval Framework
Supabase Eval
An AI database assistant agent answering natural language queries over a Supabase database via an HTTP MCP server, grounded in a pgvector knowledge base. Measured by an automated LLM-as-judge eval pipeline scoring accuracy and security across 30 test scenarios.
About Me
I'm a Junior AI Engineer at PT Sigma Cipta Utama, currently building production LLM systems and real-time AI infrastructure. CS graduate from Telkom University with a 3.96/4.00 GPA.
My work spans LLM agent pipelines, real-time audio/video AI (WebRTC), RAG systems, and microservices. I care deeply about making AI systems that actually work in production — low latency, high reliability.
I'm pursuing a MEXT research scholarship to continue AI research in Japan, focused on affective computing and multi-modal AI systems.
Read full bioLet's build something together
Open to interesting AI/LLM projects, research collaborations, and full-time roles.