Researching the Current LLM Memory Problem
My first research on the LLM memory issue.
AI engineering journal
Deep dives on retrieval systems, evaluation loops, and execution tradeoffs that matter in production.
Process
A lean delivery rhythm designed for shipping useful AI products without hype.
Discover
Map business constraints, user risk, and delivery goals before model choices.
Build
Implement the smallest robust architecture that can ship and be observed.
Measure
Track quality, latency, and reliability signals to guide iterative improvements.
Latest notes
Decision-focused notes on architecture, tradeoffs, and delivery patterns.
My first research on the LLM memory issue.
A practical architecture breakdown of my first production website: React frontend, secured API layer, Supabase auth/data, and AI-assisted grading workflows.
Builds
Selected implementations with clear scope and practical outcomes.
Built the first production website for IELTS writing practice with a practical architecture focused on reliability and fast iteration.
Shipped a production IELTS writing practice platform with secure auth, AI-assisted feedback, and stable API integrations.