Website:
goedmo.com
Job details:
About EDMO
EDMO is building the next generation of conversational intelligence for higher education. Our platform empowers universities and EdTech institutions with production-grade agentic AI - enabling intelligent student advisors, enrollment agents, academic support workflows, and enterprise automation. We work with forward-thinking institutions (including flagship US university partners) and are scaling fast. If you love designing AI systems that actually run in production, this role is for you.
Role Overview
As the Agentic AI Architect at EDMO, you will own the end-to-end technical design and architecture of our agentic AI systems. You will be the bridge between research-grade AI ideas and production-hardened implementations defining how LLM pipelines, multi-agent orchestration, memory systems, and tool-use come together into a coherent, scalable platform.
This is not a research or prototyping role. You will design systems that run at enterprise scale, handle real student interactions, integrate with university SIS/CRM platforms, and meet compliance standards (FERPA, SOC 2). Equally important, you will be able to communicate complex architecture decisions clearly to engineers, product stakeholders, and enterprise clients.
What You'll Own
System Architecture & Design :
- Define the end-to-end agentic AI architecture LLM orchestration layers, agent coordination mechanisms, memory management, tool-calling patterns, and multi-agent workflows
- Design multi-agent orchestration patterns including hierarchical agents, supervisor-worker topologies, plan-and-execute strategies, and agent-to-agent (A2A) communication
- Architect RAG (Retrieval-Augmented Generation) pipelines, hybrid search, and knowledge management systems tailored to higher-education content (catalogs, policies, student data)
- Establish architectural standards for prompt engineering, context-window management, token budgeting, and evaluation frameworks
- Design for non-functional requirements : latency targets, throughput, reliability, observability, cost controls, and scaling guardrails
Production Engineering Standards
- Build production-grade systems with structured error handling, retry logic, circuit breakers, and fallback execution paths
- Define DevSecOps architecture and deployment patterns across multi-cloud environments.
- Implement human-in-the-loop (HITL) checkpoints for sensitive decision points within student journeys
- Establish monitoring, observability, and audit logging frameworks for compliance, traceability, and agent behavior analysis
- Design secure AI architectures threat modeling, adversarial prompt defenses, and data privacy controls aligned with FERPA and SOC 2
Client & Team Communication
- Produce clear architecture documentation : reference diagrams, integration patterns, decision records, ADRs, and runbooks for production operations
- Communicate architectural trade-offs to both engineering teams and non-technical enterprise clients in a structured, accessible way
- Lead architecture and design reviews, mentoring engineers on agentic patterns and production best practices
- Serve as the technical point of contact for enterprise client integrations translating institutional requirements into AI system design
Research & Innovation
- Stay current with the latest agent frameworks (LangGraph, CrewAI, AutoGen, Google ADK), LLM advancements, and inference optimization techniques
- Build prototypes and proof-of-concepts to validate architectural approaches before committing to production builds
- Drive the technology roadmap for EDMO's agentic platform identifying where autonomous agents can replace manual workflows with measurable ROI
Core Technical Requirements
You'll Be a Great Fit If You Have :
- 5+ years of hands-on software engineering, with 3+ years specifically in production AI/ML systems
- Demonstrated experience designing and shipping multi-agent AI systems in production (not just prototypes or demos)
- Deep expertise with LLM orchestration frameworks : LangChain, LangGraph (preferred), and/or AutoGen, CrewAI, Google ADK
- Strong command of Python and backend engineering fundamentals - APIs, microservices, async patterns
- Hands-on experience with RAG architectures, vector databases, and semantic search
- Proficiency with cloud platforms (AWS, GCP, or Azure) and containerized deployments (Docker, Kubernetes)
Architecture & Systems Thinking
- Ability to design systems addressing latency, scalability, reliability, and cost simultaneously not just "it works on my local machine"
- Experience with agent memory architectures : episodic, semantic, procedural and state management in stateful multi-agent graphs
- Familiarity with LLM inference optimization quantization, model routing, caching, context compression
- Understanding of API gateway design, service boundaries, versioning, and governance
Communication & Leadership
- Ability to produce clear architecture diagrams and technical documentation accessible to both engineers and non-technical stakeholders
- Experience presenting architecture to enterprise clients or senior leadership
- Track record of leading technical design reviews and influencing architecture decisions without direct authority
Nice To Have
- Experience with voice AI / telephony
- EDMO has voice agent modules for student engagement
- Salesforce / CRM integration experience
- University CRM and SIS integrations are core to our enterprise deployments
- Higher education domain knowledge
- Understanding of admissions, advising, and enrollment workflows accelerates delivery
- FERPA / SOC 2 compliance experience
- Required for US university enterprise clients
- Familiarity with n8n or workflow automation tools
- Used in select automation workflows (but production AI agents are primary)
- Experience with Databricks / MongoDB for AI data pipelines
- Core to EDMO's data layer
- Google Gemini / Vertex AI experience
- Part of our multi-LLM strategy
What You Won't Be Doing
This role is explicitly not for :
- Research-only, notebook-bound, or demo-focused ML work
- Low-code/no-code automation tool configuration (Zapier, Make, etc.)
- "Prompt-only" AI work without systems thinking
- Architecture decks that never get built
How We Work
- Tech Stack : Python, LangGraph, LangChain, Google Gemini / Claude / OpenAI (multi-LLM), MongoDB, Databricks, Salesforce integrations,
- Deployment : Production-grade microservices, containerized on cloud with CI/CD pipelines
- Compliance Environment : FERPA, SOC 2, VPAT security and privacy are first-class concerns
- Team : Small, high-caliber engineering team. Architects write architecture AND contribute to code
- Clients : US-based universities and EdTech platforms at enterprise scale
(ref:hirist.tech)
Click on Apply to know more.