Senior AI Architect (AI-Native App Development)
Responsibilities & Duties
Role Overview
We are seeking a Senior AI Architect to lead application development using AI-native tools like Lovable.dev and Claude. You’ll rapidly prototype production-ready React/Supabase apps through natural language prompting while stepping in with Python and TypeScript when AI-generated code needs human refinement. The ideal candidate combines deep prompt engineering expertise with strong full-stack fundamentals to bridge the gap between AI capabilities and production-ready systems
Core Responsibilities
-
Rapid Prototyping & Production Build: Use Lovable to generate production-ready React/Supabase applications from natural language prompts, rapidly iterating on UI/UX and database schemas.
-
LLM Strategy & Integration (Claude): Serve as the subject matter expert on Anthropic’s Claude models (Sonnet 3.5, Opus). Design prompts and system instructions that optimize Claude’s reasoning for our specific business logic
-
Hybrid Development: Recognize the limits of AI generation. Step in with Scripting (Python/TypeScript) to handle complex edge cases, backend logic, and API integrations that AI tools cannot yet perfect on their own.
-
System Architecture: Design the data flow between our frontend apps (Lovable- generated), our backend databases (Supabase/PostgreSQL), and our AI Agents. Ensure security, scalability, and maintainability of AI-generated codebases.
-
Workflow Optimization: Build internal tools and scripts that automate the “last mile” of development, connecting our AI apps to external APIs, CI/CD pipelines, and cloud infrastructure.
-
Bring new perspectives that challenge assumptions and add original thinking.
-
Ensure approaches align with organizational goals before execution.
Technical Requirements
1. AI & LLM Expertise (The Brains)
-
Deep experience with Claude: Proven ability to use Claude 3.5 Sonnet/Opus for code generation and complex reasoning tasks.
-
Prompt Engineering: Mastery of “System Prompts” and context window management. You know how to structure a prompt to get clean, bug-free JSON or code output from an LLM.
-
Agentic Workflows: Understanding of how to chain LLM calls together to solve multi- step problems.
2. Application Development (The Build)
-
Lovable / V0 / Cursor Mastery: Hands-on experience with AI-native build tools. You specifically know how to drive Lovable.dev to create complex, multi-page applications with authentication and database connections.
-
Full Stack Fundamentals: Even though AI writes the code, you must understand React, Tailwind CSS, and Node.js to debug and refactor what the AI produces.
-
Database Design: Proficiency with Supabase or PostgreSQL. You can define schemas, Row Level Security (RLS) policies, and relationships that an AI tool can implement.
3. Scripting & Automation (The Glue)
-
Proficient in Python: For backend data processing, RAG (Retrieval Augmented Generation) pipelines, and server-side logic.
-
TypeScript/JavaScript: For client-side logic customization within generated apps.
-
API Orchestration: Experience connecting REST and GraphQL APIs.
Key Tech Stack
-
AI Models: Anthropic Claude 3.5 Sonnet (Primary), GPT-4o.
-
Gen-AI Tools: Lovable.dev (Required), Cursor IDE, V0.
-
Backend: Supabase, PostgreSQL, Python (FastAPI/Flask).
-
Frontend: React, TypeScript, Tailwind CSS.