This is a hands-on engineering role. You are writing code daily, iterating quickly, and working directly with modern AI tooling. You are expected to understand how LLMs behave in production, not just how they work in theory.
Design, build, and deploy production-grade AI agents and end-to-end agentic workflows that solve real business problems across ACQ Vantage
Integrate LLMs with internal systems, APIs, and data sources, ensuring reliability, performance, and clean abstractions
Collaborate with product and engineering teams to prioritize, ship, and iterate on AI features quickly
Own and improve RAG pipelines across multiple Pinecone namespaces, including chunking strategy, embedding model selection, hybrid retrieval, and reranking
Build and maintain an evaluation framework, including golden datasets, automated quality scoring, retrieval metrics, latency benchmarks, and regression detection
Optimize model routing and tiering to improve unit economics while maintaining output quality
Instrument the AI layer for observability, including cost-per-request, token usage, quality signals, and anomaly detection
Qualification
Required
7+ years shipping production software systems (distributed backends, APIs, deployment pipelines, monitoring)
2+ years building production RAG systems using vector databases (Pinecone, Qdrant, FAISS, or Weaviate), including embedding strategies, index management, and retrieval tuning
Built and deployed AI agents or multi-step LLM workflows in production, including tool use, orchestration, and system integrations
Built or contributed to an evaluation framework for an LLM-based product (retrieval quality measurement, regression detection, model-switching decisions based on data)
Reduced LLM API costs in production through model routing, caching, token management, or architectural improvements
Worked across multiple LLM providers (OpenAI, Anthropic, or equivalent) and understands tradeoffs in prompt behavior, token economics, and failure modes
Comfortable in both TypeScript and Python (our stack uses both)
Production AI agents are deployed and actively used within ACQ Vantage
New AI-driven features move from concept to production in weeks, not months
Agent performance improves over time through structured testing and iteration
AI systems operate reliably with minimal failure or manual intervention
Engineering output translates directly into measurable business impact
