Sayari is a venture-backed global corporate data provider and commercial intelligence platform. They are seeking a Senior AI/ML Engineer to architect multi-agent orchestration systems that enhance how analysts interact with global risk data, ensuring the deployment of trustworthy AI interactions.
Responsibilities:
- Architect and deploy complex, multi-agent orchestration workflows end-to-end using frameworks like Langchain
- Build and maintain a robust production evaluation framework for both online and offline LLM performance, ensuring our systems are accurate, reliable, and defensible
- Develop and optimize orchestration and routing protocols to accelerate AI development and unblock cross-functional teams building on difficult problems
- Serve as a technical mentor and thought leader for teams, advising on prompt engineering, workflow design, and the critical trade-offs between classical ML and agentic approaches
- Collaborate directly with the AI Innovation Lab Manager and SVP of Engineering to align technical execution with Q2 product commitments
- Independently solve complex engineering hurdles in the agentic stack, moving from POC to a deployed MVP within a 90-day window
Requirements:
- 5+ years of engineering experience
- 2+ years of experience with full lifecycle machine learning activities
- 2+ years of agentic AI experience
- Production AI/ML Expertise: You have moved beyond notebooks and have a track record of deploying multi-agent workflows at scale that incorporate reasoning loops (CoT, ReAct) and tool use (RAG, API calls)
- Orchestration & Routing: Mastery of Langchain or similar frameworks, with the ability to build clean, maintainable, and observable agentic workflows
- Cloud Engineering: Solid background in cloud infrastructure (AWS/GCP/Azure) with the ability to write production-ready code
- ML Logic & Discernment: A solid background in classical ML (classification, regression) used to determine when to build agentic workflows versus utilizing traditional statistical models
- Decisive Execution: A builder mindset with the ability to work independently, navigate ambiguity, and pivot quickly as priorities evolve
- Deep experience with Orchestration/Routing, ReAct, and Chain of Thought
- Background in RAG, GraphRAG, and evaluating workflows using LLM-as-a-judge vs. deterministic scoring
- Experience in GCP