The property management industry is ripe for disruption, and the perfect vertical for natural language automation.  Haven is building an autonomous workforce for the property management industry. Our suite of AI agents connect to property data and integrate with a property manager’s software to automate manual, rote tasks on top of the systems that they already use today.  Our AI agents are already live across 10,000+ residential properties, and we’re on track to double before the end of the year. We’re backed by top VCs such as YCombinator, and incredible angels like the founder of Salesloft.  Our Co-founders previously built and sold a vertical SaaS company in less than 2 years, disrupting another ossified industry (K-12 admissions). The Problem We're Solving: The US rental industry generates $1+ Trillion per year, most of which is administered by property management companies that run on archaic Property Management Systems (PMS). These Legacy PMS generate Billions in revenue, yet they remain slow, incredibly complicated to use, and painfully manual. These monolithic systems automate only a small part of a property manager’s work, forcing companies to scale by hiring additional, often less experienced staff for labor-intensive tasks. This approach adds inefficiency and cost without truly solving the underlying problem. As a result, the property management industry suffers from razor-thin margins, a 70% employee turnover rate, and difficulty scaling efficiently. Our Solution: We’re leveraging the recent advancements in LLMs to automate the industry’s most arduous manual workflows — and wedging ourselves into one of the largest vertical SaaS opportunities that exist today. Our first autonomous worker is automating the entire maintenance coordination process. This means answering the phone when tenants have a maintenance issue, diagnosing the issue, dispatching a technician, and then following up with tenants to make sure the issue was resolved. This means property managers can scale their operations without growing headcount, while providing better, faster service to their tenants. Your Role: You'll work closely with the founders to discover, spec, and build automations that transform how property management companies operate. As a founding engineer, you'll tackle fascinating technical challenges at the intersection of LLMs and legacy software systems, helping us create truly autonomous workers that can handle complex human workflows. Complex Architectural Puzzles: Navigate challenges in natural language interfaces, model evaluation complexities, customizing models to diverse client-specific needs, and multi-modal delivery (i.e., SMS, Voice, and embedded chat widget). Domain-Specific Models: Fine-tune and evaluate models to handle the nuanced requirements of property management, including maintenance diagnosis, vendor coordination, and tenant communication. Model Evaluations and Algorithms: Conceptualize and implement multi-step conversational workflows, retrieval augmented generation (RAG), search, and QA algorithms. Engineering Large-scale LLM Systems: Ensure optimal performance while engineering and assessing vast interconnected LLM systems. Data Harvesting and Structuring: Design and build pipelines that can ingest, structure, and retrieve diverse property management data to augment our LLMs with domain knowledge. High-Performance Infrastructure: Optimize for quick response times and efficient model inference while maintaining reliability at scale. How You'll Work: Project Ownership: Own projects from start to finish and architect solutions on your own with limited guidance. This could look like talking to customers, scoping out a solution, and then building, testing, and documenting it — all within a few days! Full Stack Development: Work across the entire tech stack, writing backend, frontend, and LLM system code. Complete Autonomy: You’ll be on the cutting edge of a developing technology space. Experimentation and rapid iteration aren’t just encouraged, they’re expected! This is a unique opportunity to shape the future of the housing industry while working with the latest advancements in AI. You'll be building something truly novel – not just another API wrapper around ChatGPT, but sophisticated systems that can reliably automate complex human workflows. What We are Looking for Experience: At least 2 years of hands-on experience with software engineering. Alternatively, 2+ years working more generally in AI/ML. Versatility: You are comfortable working at every level of the stack; from writing backend and frontend code, to designing LLM chains, to implementing necessary tools or processes. Autonomy: You don’t require heavy direction, enjoy open-ended problems, and thrive in the face of uncertainty – typical elements of a startup environment. Decision-Making Skills: Ability to make quick technical decisions and clearly communicate your decision-making process. Communication Skills: You’re clear and concise in explaining your thought process and in enumerating the tradeoffs made in your work. Relocation: Open to relocating to NYC. Passion for LLMs: You keep up with the latest developments in AI/LLMs in your spare time and are eager to apply these to solve real-world problems. Bonus: Familiarity with emerging prompting techniques and familiarity building agents over different modalities (voice, email, SMS). Bonus: Experience in building AI agents (e.g., with langchain's/llamaindex’s agents, memory, and chain modules). We want to see what you’ve built! Technology Stack Frontend Vue JS Vite Tailwind Backend NodeJS Express Typescript MongoDB ML GPT, Claude, and Llama family of models Embeddings