aicancode

DevBlog: build a secure blog REST API (Java, Python or TypeScript) from requirements to monitoring, with AI as your pair programmer. main = starter, solution = reference with per-phase tags.

FunnelLens: a data analyst case study on a messy e-commerce funnel (pandas, DuckDB SQL or tidyverse): data contract, confounder-adjusted tests, three-track parity, CI, a Dockerised dashboard with a guard-railed AI narrator, monitoring and a post-mortem. main = starter, solution = reference with per-phase tags.

RateGuard: build a distributed rate limiter (Java, Python or TypeScript) with atomic Redis Lua scripts, per-tenant fail modes and an AI policy advisor, from requirements to monitoring, with AI as your pair programmer. main = starter, solution = reference with per-phase tags.

ShopFlow: build an e-commerce backend (Java, Python or TypeScript) with idempotent checkout, signed payment webhooks and AI search, from requirements to monitoring, with AI as your pair programmer. main = starter, solution = reference with per-phase tags.

ChurnCast: an end-to-end data science case study on subscription churn (pandas + scikit-learn, DuckDB SQL or tidymodels): point-in-time features, leakage tests, calibrated models, three-track parity, CI with an evaluation gate, a Dockerised API with a guard-railed AI explainer, monitoring and retraining. main = starter, solution = reference.

Pipewright: an end-to-end data pipeline for a D2C brand's CEO dashboard. Python ingestion (API + file drops) into a lake, a versioned data contract, dbt on DuckDB or PySpark with cross-track parity, a data-quality publish gate, Airflow, Streamlit, Docker, incremental models, backfills and SLA alerts. main = starter, solution = reference.

SnapLink: build a production-style URL shortener (Java, Python or TypeScript) from requirements to monitoring, with AI as your pair programmer. main = starter, solution = reference with per-phase tags.

StudyBuddy: build a RAG study assistant (Java, Python or TypeScript) that answers from your own notes with page citations, says "I don't know" when it should and is evaluated in CI, with pgvector + full-text hybrid search, from requirements to monitoring, with AI as your pair programmer. main = starter, solution = reference with per-phase tags.