AI Developer

We’re hiring an AI Developer to help us build and scale production-grade AI agents. This role focuses on agentic systems that use RAG, vector databases, and structured tool use to deliver reliable outcomes. You’ll work end-to-end—from architecture to deployment—on AI systems that go well beyond prompt experiments.

What You’ll Do

• Build and maintain AI agents capable of reasoning, planning, and executing tasks
• Design and implement Retrieval-Augmented Generation (RAG) pipelines
• Apply effective data chunking, embedding, and retrieval strategies
• Integrate vector databases into agent workflows
• Use and extend AI agent frameworks (LangChain, LlamaIndex, AutoGen, CrewAI, etc.)
• Connect agents to APIs, internal systems, and external tools
• Optimize prompts, agent logic, and memory for reliability and performance
• Deploy, monitor, and iterate on AI systems in production
• Collaborate closely with product and engineering teams

Requirements

• Strong proficiency in Python and/or JavaScript/TypeScript

• Hands-on experience building AI agents or LLM-powered applications
• Practical experience with RAG architectures
• Solid understanding of vector databases (e.g., Pinecone, Weaviate, FAISS, Chroma)
• Experience with data chunking, embeddings, and retrieval tuning
• Familiarity with AI agent frameworks such as LangChain, LlamaIndex, AutoGen, or similar
• Experience integrating LLMs with tools, APIs, and structured data
• Ability to reason about and debug agent behavior in real-world scenarios

Nice to Have

• Experience with multi-agent systems
• Production deployment experience on AWS, GCP, or Azure
• Familiarity with LLM evaluation, monitoring, and observability
• Background in backend engineering or distributed systems
• Experience working in startup or fast-paced product environments

What Success Looks Like

• You ship reliable, production-ready agents
• RAG systems you build retrieve relevant, high-quality context
• Agents behave predictably and improve over time
• You proactively identify and fix failure modes
• You balance speed with system robustness

Originally posted on Himalayas