Build logs from the things that went wrong.
Open the journalHands-on RAG practice — loaders, splitters, ChromaDB vector stores and LCEL runnables.
Tracking prompts, latency and token costs in LangSmith — and tuning a PDF RAG pipeline over 4 iterations.
Stateful LangGraph workflows — from parallel essay graders to equation solvers and sentiment routers.
Custom FastMCP servers, emailmd API tools and an interactive Streamlit chatbot client.
From clicking around the AWS console to Terraform modules, S3 state locking, Ansible playbooks and Vault secrets.
FIFO batch allocation, concurrency-safe PostgreSQL transactions, ReportLab invoices and instant Telegram alerts.
How the LangGraph editorial pipeline filters technical noise into sharp 5-point developer briefs.
An async pipeline that crawls the web, cleans pages, drafts reports, audits facts and indexes vectors in Pinecone.
Leading a team of five on a dual-role web and React Native app with dynamic watermarking and phase-based unlocking.
The story of MockVue: AI feedback, live code execution and analytics, and every wrong turn on the way.
The build log behind AI Data Analyst: turning raw datasets into LLM-generated insights, charts and summaries.
Docker, Jenkins, Terraform and AWS EC2 — what it actually took to ship a Flask app end to end.