About The Role The role involves architecting and scaling Large Language Model systems that move beyond experimental notebooks into robust production environments. This position focuses on the intersection of generative AI and software engineering, requiring a deep understanding of how to optimize model performance, manage context windows, and ensure output reliability. The team builds the core infrastructure that powers intelligent applications, focusing on retrieval-augmented generation (RAG), agentic reasoning loops, and high-throughput inference pipelines. This role is critical for transforming raw foundation models into specialized, high-accuracy tools that solve complex business logic challenges.

Key Responsibilities

• Architect and deploy production-grade RAG pipelines using LangChain or LlamaIndex, incorporating advanced retrieval techniques like hybrid search and reranking.
• Implement and maintain vector database infrastructure using Pinecone, Milvus, or Weaviate to handle multi-million document embeddings with low latency.
• Develop automated LLM evaluation suites to measure hallucination rates, groundedness, and relevance using frameworks like RAGAS or custom LLM-as-a-judge patterns.
• Optimize model inference costs and latency through techniques such as prompt caching, quantization, and fine-tuning with LoRA/QLoRA on domain-specific datasets.
• Build and integrate agentic workflows that leverage tool-calling and multi-step reasoning to automate complex analytical tasks.
• Collaborate with data engineers to build robust ETL pipelines that transform unstructured data into high-quality training and retrieval sets.

What We Are Looking For

• 3-6 years of experience in software engineering, with at least 1.5 years dedicated to deploying LLM-based applications in production.
• Expert-level Python proficiency, including experience with asynchronous programming and building high-performance APIs (FastAPI/Flask).
• Demonstrated experience with vector databases and a deep understanding of embedding models and semantic similarity metrics.
• Hands-on experience with LLM orchestration frameworks and a solid grasp of prompt engineering best practices and versioning.
• B.S. or M.S. in Computer Science, Data Science, or a related technical field.
• Bonus: Experience with fine-tuning open-source models (Llama 3, Mistral), familiarity with vLLM/TGI for serving, or contributions to AI open-source projects.