33 prompts
Outline and code a simple RAG prototype in Python that answers questions from [documents]. Cover loading and chunking documents, creating embeddings, storing them in [vector store], retrieving relevant chunks, building the prompt, and citing sources in answers.
Answers questions from your own documents using embeddings.
Review this PostgreSQL schema. Check normalisation, data types, constraints, foreign keys, indexes for likely queries ([queries]), naming consistency, and future scaling issues. Suggest improvements with migration SQL. Schema: [paste schema]
Reviews a database schema for performance and integrity issues.
I want to build [app idea] for [users]. Suggest a simple tech stack for a solo developer, the core MVP features, a data model, main pages/screens, API endpoints, auth approach, hosting, and a 4-week build plan. Flag what to leave out of the MVP.
Turns an app idea into an MVP plan with stack, data model and milestones.
Analyze this data and tell me what decisions it supports. Summarise the key trends, anomalies, and comparisons, then give 3 recommended actions with the evidence for each and what additional data would increase confidence. Data: [paste data]
Turns a data export into insights and recommended decisions.
Analyze this churn data and cancellation reasons. Identify the top reasons customers leave, which segments churn most, early warning signals, and 5 retention actions ranked by impact. Data: [paste data]
Finds patterns in why customers leave and what to do.
Analyze these survey responses and build 3 marketing personas. For each: name, demographics, goals, pains, buying triggers, objections, preferred channels, and 2 real quotes from the data. Note how many respondents fit each persona. Responses: [paste data]
Builds evidence-based personas from real survey answers.
I'm interested in researching [broad topic] for my [degree] thesis in [field]. Help me narrow it down: suggest 5 focused research questions, evaluate each for originality, feasibility and available data, recommend the strongest one, and outline possible methods and key literature to start with.
Turns a broad interest into a focused, researchable question.
You are a machine learning engineer. Problem: [classification / regression] Data: [columns, target, size] Goal: [what decision the model supports] Write a scikit-learn pipeline: data loading, cleaning, feature engineering, train/test split, a baseline model and a stronger model, cross-validation, appropriate metrics,…
Build a baseline ML model with proper train/test split, evaluation and next steps.
You are an engineer who builds production AI features. Feature idea: [e.g. summarise support tickets, auto-tag products] Stack: [details] Data involved: [types, sensitivity] Design: the prompt and system instructions, input/output format (JSON schema), handling of errors and timeouts, cost and latency considerations,…
Plan and build an AI feature, such as summarising, classifying or a chat assistant, with good prompts and guardrails.
You are a cloud solutions architect. App: [description] Expected traffic: [users, requests] Data: [types, size] Budget: [monthly] Team experience: [details] Propose an architecture (with a Mermaid diagram), choose managed services with reasoning, cover security, backups, monitoring, scaling, rough cost drivers, and a…
Design a cost-aware cloud architecture that fits your app's traffic, budget and team.
You are a performance-focused backend engineer. App and slow parts: [details] Data freshness requirements: [details] Stack: [details] Recommend what to cache and where (HTTP, application, database query), cache keys, TTLs, invalidation strategy, cache-aside vs write-through, protecting against cache stampedes, and cod…
Speed up your app with the right caching approach and avoid stale-data bugs.
You are a Node.js backend engineer focused on security. Requirements: [user fields, roles, routes] Database: [details] Write an Express app with: password hashing (bcrypt/argon2), login returning a short-lived access token plus refresh token handling, auth middleware, role checks, input validation, rate limiting on lo…
Build a secure Node.js/Express API with sign-up, login, protected routes and token handling.