Senior AI Engineer

Role purpose

Key responsibilities

• Architect production LLM applications, multi-agent systems, and RAG pipelines; select models and balance cost, latency, and quality.
• Design context engineering for retrieval, tools, memory, and grounding in project, inventory, pricing, and CRM data.
• Build evaluations, red-teaming, and guardrails for safety, factual grounding, and hallucination control.
• Set standards, run design reviews, mentor engineers, and lead integrations with Salesforce, ERP, and buyer/investor portals.
• Evaluate emerging agentic, computer-use, and voice AI capabilities; partner on data and vector infrastructure.
• Present roadmap, risk, and ROI to leadership and contribute to cross-functional planning.

Technology stack and AI expertise

• Languages: Python and TypeScript/Node.js. LLM APIs include Anthropic Claude, OpenAI, and Azure OpenAI.
• Frameworks include LangChain, LangGraph, LlamaIndex, Anthropic Agent SDK, and MCP; retrieval options include Pinecone, pgvector, Weaviate, and Azure AI Search.
• Backend options include FastAPI or Node/Express and REST/GraphQL; cloud stack includes Azure, Docker, Kubernetes/AKS, and CI/CD.
• Data and evaluation tools include PostgreSQL, Azure Blob/Data Lake, Airflow/ADF, LangSmith, Langfuse, PromptFoo, and custom evaluation harnesses.
• Relevant practices include agentic AI, enterprise RAG, hybrid search, prompt/context engineering, structured outputs, tool calling, model evaluation, guardrails, and cost/latency optimization.
• Fine-tuning/LoRA, voice AI, and computer-use agents are listed as areas of familiarity or exploration.

Must-have qualifications

• Strong software engineering background in Python and/or TypeScript/Node; experience shipping LLM applications to production.
• Deep RAG, embeddings, vector database, and agent-framework knowledge; experience with Anthropic, OpenAI, or equivalent APIs.
• Solid system design, API design, and cloud infrastructure knowledge; Azure preferred.
• Technical leadership or mentoring experience.

Nice-to-have qualifications

• LLMOps evaluation, observability, or guardrail experience; real-estate, PropTech, or enterprise Salesforce/ERP integration experience.
• Familiarity with fine-tuning or model customization; exposure to voice AI or computer-use/browser agents.

Business domain

Application process

• Candidates are asked to complete assessments and interviews independently, without generative AI assistance.