**Location:** Remote (European timezone overlap required) **Engagement:** Part-time to start, potential long-term **Rate:** Discuss during screening --- ## About the Project We're building an intelligent, conversation-based interface for the smart-home industry. Our platform unifies IoT device data through natural language – users ask questions, and our AI agent retrieves documentation, analyzes device telemetry, and provides actionable answers. The backend exists and runs in production. We need a skilled developer to take our LLM integration to the next level: better retrieval, smarter agents, and more robust multi-turn conversations. --- ## Tech Stack - Python 3.12, FastAPI, async/await - LangChain / LangGraph (state machine agents) - Multi-provider LLM integration (OpenAI, Anthropic, Fireworks) - PostgreSQL via Supabase (considering pgvector) - IoT protocols: KNX, Modbus, ThingsBoard - Kubernetes deployment --- ## What You'll Work On - Enhance RAG pipeline with semantic search (vector embeddings) - Improve multi-turn conversation handling - Optimize agent routing and tool selection - Add new capabilities (device control, automation suggestions) --- ## What Matters to Us We don't care about years of experience or job titles. We care about: - **You've built something real with LLMs** – Show us, tell us about it - **Clean async Python** – You understand when to use await, how to structure services - **You can communicate clearly** – In English, in writing, on video calls - **You take ownership** – You figure things out, ask good questions, deliver **Bonus points:** - LangChain/LangGraph experience - Vector databases (pgvector, Pinecone, Weaviate) - IoT/home automation background --- ## Our Process 1. **Application review** – We read every response carefully 2. **Video call (30 min)** – Get to know each other, discuss your work 3. **Technical discussion** – Walk through architecture, your approach 4. **Paid trial task** – Small real-world task to confirm fit 5. **Ongoing collaboration** We believe in conversation first. No coding tests before we talk. --- ## To Apply Please include (generic applications will be ignored): 1. **Your background** – Who are you? What excites you about AI? 2. **A project you're proud of** – Describe something you built with LLMs or AI. What problem did you solve? What was hard? 3. **Quick technical question** – In a RAG system, when would you use hybrid search (keyword + semantic) vs pure semantic search? 4. **Availability** – Hours per week and your timezone 5. **Rate expectation** – Hourly or monthly 6. **Show your work** – GitHub, portfolio, blog, or anything that demonstrates your skills --- ## About Us Small European startup in the smart-home/IoT space. Technical founder, quality-focused, async-friendly. We value clean code, clear thinking, and people who ship.