About The Role
The role is for someone who has moved beyond basic prompting and understands what it takes to build production-grade AI systems: robust RAG pipelines, agentic workflows, fine-tuning pipelines, and systematic evaluation frameworks.
The engineering team owns complex pieces of a high-scale AI platform, working directly with applied scientists and backend engineers to deploy performant LLM applications.
Key Responsibilities
• Design and implement production RAG pipelines using LangChain, LlamaIndex, or custom Python architectures
• Build and optimize vector database integrations such as Pinecone, Weaviate, or pgvector for low-latency semantic search
• Develop systematic LLM evaluation frameworks including benchmark suites, LLM-as-judge pipelines, and automated regression testing
• Execute parameter-efficient fine-tuning pipelines using LoRA and QLoRA on specialized domain datasets
• Write observable, tested, and well-documented Python code; participate actively in architecture reviews and deployment pipelines
What We Are Looking For
• 3 to 6 years of software engineering experience, with at least 2 years specifically focused on building LLM applications in production
• Deep familiarity with LLM orchestration frameworks, prompt engineering best practices, and context window optimization
• Strong proficiency in Python, async programming, REST API design, and cloud infrastructure integration
• Solid understanding of embedding models, vector spaces, tokenization, and model quantization techniques
• Bonus: Experience with custom CUDA kernels, open-source model deployment via vLLM or TGI, and published contributions to AI open-source projects