AI Engineer

CompanyEighty Days is building a new way to discover the world centred on taste, trust, and personal context. Today, travel discovery is fragmented across social media, search, and private conversations. We are creating a product that brings this together into a structured, map-based experience where recommendations are attributed, meaningful, and built over time.Our focus is not just on helping people plan trips, but on becoming the place where they continuously save, organise, and revisit places that matter.Website: https://eightydays.ai/What you'll doDesign, build, and own LLM-powered product features (chat and conversational modules, content understanding, recommendations) end to end on our NestJS/TypeScript backend: from product requirement, to production, to iteration based on real user behavior.Design and own the LLM interface layer: orchestration, tool calling, memory, evaluation, and the glue that makes agents and multi-step pipelines reliable in production.Work fluently across many LLMs (proprietary and open-source), choosing and routing the right model for each job based on quality, latency, and cost.Develop and maintain RAG systems: embedding generation, vector search (Qdrant), hybrid retrieval with Elasticsearch and MongoDB, and grounding LLM responses in first-party data.Engineer for production performance: SSE streaming, prompt caching, token budgets, Redis-backed session state, SQS-based async processing, and observability.Build automated test suites and internal LLM evaluation tooling: regression tests for prompt and model changes, eval datasets curated from real user queries, and metrics for relevance, response quality, and groundedness.Translate business and product insights into working AI features, collaborating closely with the founder, backend engineers, and product.What we're looking for1.5+ years of software engineering experience with strong TypeScript/Node.js skills (NestJS preferred).Shipped at least one LLM-powered feature to production (chatbot, assistant, RAG system, content extraction pipeline, etc.), not just prototypes.Hands-on experience with a major LLM API (Gemini, OpenAI, or Anthropic): prompt design, structured outputs, function/tool calling, and streaming.Working knowledge of embeddings, vector search (Qdrant, Pinecone, pgvector, etc.), and RAG patterns.Production experience with MongoDB, Elasticsearch, or Redis.Strong product sense: able to judge what makes an AI response genuinely useful and to balance quality, latency, and cost trade-offs.Ownership mindset: can own a module end to end, from ambiguous requirement to shipped feature, with minimal oversight.Familiarity with AWS fundamentals (EC2/Beanstalk, S3, SQS).Bachelor's in Computer Science, Engineering, or related field, or equivalent practical experience.Nice to haveExperience building LLM evaluation pipelines or working with eval frameworks.Exposure to agent frameworks (LangGraph, LangChain) or Model Context Protocol (MCP). Note: we build most pipelines directly on model APIs, so framework knowledge is a bonus, not a requirement.Experience with prompt caching, model routing, or LLM cost optimization at scale.Experience with messaging platform APIs (WhatsApp Cloud API, Instagram Graph API).Prior experience in a fast-paced, AI-driven startup environment.