Agentic AI Architect (Senior)

Role: Agentic AI ArchitectLocation: Remote (US)Overview

Client delivers data and AI-powered solutions that meet the needs of our customers.

The team focuses on improving operational efficiency across the payments ecosystem. They are expanding with agentic AI, LLM-driven workflows, and autonomous operational agents.

We are seeking an Agentic AI Architect to architect, build, and scale highly scalable multi-agent systems, distributed reasoning engines, and intelligent automation platforms.

This role begins as a hands-on individual contributor (IC) responsible for designing and implementing the highly scalable agentic AI products end-to-end.

Role ResponsibilitiesPlatform Architecture & System Design

• Architect multi-agent LLM systems using LangGraph, LangChain, AutoGen, Strands, Semantic Kernel, Haystack Agents, HuggingFace Agents, and OpenAI Agent SDK.
• Design stateful, deterministic, and parallel agent workflows with guardrails, fallback logic, and structured error recovery.
• Define workflows for summarization, extraction, diagnostics, reasoning, grounding, and classification.
• Build evaluation frameworks for accuracy, grounding quality, hallucination detection, latency, and agent performance.
• Lead architecture of secure, distributed, event-driven microservices supporting agent collaboration and tool calling.

Agentic AI Engineering & Orchestration

• Build modular agent ecosystems (classification, routing, reconciliation, anomaly detection, knowledge agents).
• Implement advanced memory systems: short-term, long-term, working, episodic, semantic, vector, and graph memory.
• Develop agent orchestration, message-passing, tool execution policies, and model routing flows.
• Use LangGraph to define state machines, execution graphs, and structured reasoning paths.

Model Integration & Prompt Engineering

• Integrate LLMs including GPT-4o, Claude, PaLM2, Mistral, LLaMA, and internal models.
• Build automated flows for prompt routing, variant testing, grounding validation, and chain-of-thought evaluation.
• Create reusable libraries for templates, reasoning patterns, fallback rules, and AI safety controls.

Data & System Integration

• Build ingestion pipelines for PDF filings, statements, logs, and mixed-format datasets.
• Implement backend components such as MongoDB metadata stores, Redis Streams, and internal event bus connectors.
• Integrate agentic intelligence into operational intelligence stack and backend workflows.

Production Deployment & Observability

• Deploy agentic systems on AWS EKS / Azure AKS with secure CI/CD pipelines (Azure DevOps).
• Implement full observability using OpenTelemetry, Prometheus, Signoz, Grafana, tracking:
• Latency
• Routing accuracy
• Agent-state health
• Drift & anomalies
• Build automated regression suites using golden datasets, grounding checks, and structured assertions.

Cross-Functional Collaboration

• Partner with DevOps, MLOps, platform engineering, and product teams.
• Work with OI domain SMEs to translate operational workflows into agentic AI systems.
• Collaborate with UI/API teams (React/FastAPI) to embed intelligent agentic workflows.
• Mentor teammates and, after launch, lead and scale a high-impact product engineering team.

All About YouCore Architecture & Engineering Experience

• 7-10 years in AI/ML engineering with deep experience in LLM systems, agentic architecture, or enterprise AI platforms.
• Led architecture for at least one 0→1 enterprise AI product/platform and built an enterprise-grade, safe, collaborative multi-agent system capable of automating complex workflows.
• 8–15 years’ experience in enterprise software/data engineering with 3+ years leading AI/ML/LLM platform development.
• Proven and demonstrable experience building and deploying at least one multi-agent system in production.
• Proven and demonstrable experience building and deploying at least one ML/AI product in production.
• Proven and demonstrable experience building text-to-SQL agents, including schema parsing, SQL generation, and safety validation.
• Strong foundation in distributed systems, microservices, and event-driven architecture.

Technical Skills

• Agentic Frameworks: LangGraph, LangChain, AutoGen, Strands, Semantic Kernel, Haystack Agents, HuggingFace Agents, OpenAI Agent SDK.
• LLMs: GPT-4o, Claude, Mistral, LLaMA, enterprise foundation models.
• Infra & MLOps: Kubernetes (EKS/AKS), Docker, FastAPI, Redis Streams, MongoDB, Azure DevOps.
• Evaluation: Promptfoo, grounding validation, hallucination scoring, BLEU/ROUGE, agent performance metrics.
• Programming: Python, SQL, APIs, microservices.

Memory & Reasoning Expertise

• Hands-on experience implementing:
• Short-term, long-term, working memory
• Episodic/state memory
• Semantic memory (embeddings)
• Vector memory
• Graph memory for relationship/state reasoning
• Ability to design stateful, deterministic, context-aware agent workflows.

Leadership & Collaboration

• Starts as an individual contributor, leading architecture and development hands-on.
• After launch, transitions into people leadership, building a high-impact product engineering team.
• Strong communicators across engineering, product, OI business teams, and leadership.
• Skilled in build–vs–buy evaluations, governance, and long-term platform planning.

Preferred Domain Experience

• Payments, operational intelligence, reconciliation, fraud/AML, compliance, or disputes.