SENIOR APPLIED AI ENGINEER (AGENTS) |- AI Platform

Our client building next-generation AI platform - generative AI + simulation-powered search engine that allows non-technical users to query complex datasets and automate workflows.

The Role: Senior hands-on role - own end-to-end agentic systems stack from LLM orchestration and tool-use to evaluation and production deployment. Work directly with founding and senior leadership to set technical direction and ship with product and research teams.

Key Responsibilities

Agent Development & Architecture

• Architect and build autonomous AI agents capable of querying data, answering data-based questions, and automating complex enterprise workflows for non-SQL / non-technical users
• Design multi-step reasoning, tool-use, function calling, and RAG pipelines - enabling agents to interact with data lakes, feature stores, APIs
• Implement agentic frameworks (LangGraph, AutoGen, CrewAI, or custom orchestration) with robust state management, memory, and planning
• Build evaluation harnesses for agent reliability, accuracy, safety

LLM & Generative AI Platform

• Fine-tune and deploy LLMs for enterprise-specific tasks - instruction tuning, RLHF/DPO, domain adaptation
• Optimize inference for low-latency, cost-efficient production - quantization, caching, batching, GPU/TPU scheduling
• Build retrieval systems (vector DBs - Pinecone, Weaviate, Milvus, Qdrant) for enterprise knowledge bases
• Partner with Data Engineering to ensure feature store and data lake are agent-ready

Data & Tool Integration

• Create tools and connectors for agents to interact with enterprise systems - data warehouses, lakehouses (Databricks, Snowflake), APIs
• Implement observability, logging, tracing for agentic workflows (LangSmith, Langfuse, OpenTelemetry)

Candidate Profile - Senior Level 7-12+ Years

• Proven Agent Builder: 3+ years building production LLM agents / RAG systems / AI copilots that query data and automate workflows. Portfolio of shipped agents
• LLM Stack: Deep hands-on with LLMs (GPT-4, Claude, Llama, Mistral), prompt engineering, function calling, tool-use, RAG, vector DBs, LangChain/LlamaIndex/LangGraph
• Production AI: Reliability, latency, evaluation, safety. MLOps / LLMOps
• Core Engineering: Strong Python, SQL, Spark. Databricks/Snowflake
• Enterprise AI Mindset: Simplifying complex data queries into natural language - as technical as possible, as commercial as possible