Associate, AI Agent Engineer

About this role Role Overview The Senior AI Agent Engineer brings depth of experience and sharpness of judgment to the design and delivery of production-grade AI systems at the core of a data and knowledge product business. He/She/They take ownership of architecting sophisticated multi-agent orchestrated workflows that combine GenAI and Vision AI to extract, interpret, and structure unstructured data into high-fidelity knowledge outputs that are commercialized to customers. The Senior Engineer operates with significant autonomy — making consequential architectural decisions across the GenAI-Vision AI stack, selecting the right agent orchestration patterns for each problem, and holding the line on accuracy, scalability, and reliability in customer-facing systems. He/She/They work closely with AI/ML teams, data scientists, Engineering leads, and domain experts to ensure the resulting knowledge products are not just functional but genuinely trustworthy at commercial scale. Beyond their own delivery, the Senior Engineer actively elevates the team — through design reviews, technical mentorship, and a deep investment in engineering rigor. This is a role for an engineer who combines technical craft with commercial instinct — someone who can cut through the noise of a fast-moving field, identify what actually works at scale, and build systems that hold up in the hands of paying customers. Roles & Responsibilities Architect, build and deliver complex multi-agent orchestrated workflows that integrate GenAI and Vision AI capabilities to extract and structure unstructured data at production scale Lead the technical design of end-to-end pipelines — from raw document/image ingestion through extraction, interpretation, validation, and structured knowledge output Drive the selection and integration of agentic frameworks, LLMs, vision models, and extraction tooling based on rigorous trade-off analysis (accuracy, latency, cost, maintainability) Define and implement systematic evaluation frameworks for agent outputs — measuring extraction accuracy, completeness, and consistency at the product level Build robust observability, logging, and monitoring frameworks for multi-agent systems running in production Lead efforts to address core engineering risks in knowledge extraction — hallucinations, vision model failures, edge-case document formats, accuracy degradation, and latency at scale Collaborate with AI engineering Lead (VPs) and business stakeholders to shape solution design, ensuring the AI systems built can be credibly commercialized as data/knowledge products Mentor junior engineers (analysts), conduct rigorous code and design reviews, and contribute to team-wide standards for AI engineering Required Skills & Experience Technical Skills 3–5 years of software engineering experience with at least 2–3 years of focused, hands-on work in LLM-based application development and agentic AI systems Deep expertise across multiple agentic frameworks — LangChain, LangGraph, AutoGen, CrewAI, or custom-built orchestration systems Proven experience designing and shipping multi-agent architectures in production, including tool use, memory, planning, and output validation modules Strong command of GenAI, and working understanding of various agentic solution approaches using LLMs, Embeddings, graph, MCP, etc. Strong command of Vision AI and document understanding — OCR engines, layout-aware models (LayoutLM, Donut, Pix2Struct), multimodal LLMs, image parsing, and visual grounding techniques Hands-on experience building extraction pipelines for unstructured data — PDFs, images, scanned documents, semi-structured text — and producing reliable structured outputs (JSON schemas, knowledge graphs, relational data) Strong Python skills and proficiency with AI/ML tooling including vector databases (Pinecone, Weaviate, pgvector), embedding models, and retrieval systems Solid grasp of LLM evaluation techniques — benchmark design, ground-truth const ... (truncated, view full listing at source)