Agentic AI Architect

TheAppLabb•Toronto, ON

1d•Hybrid

About The Position

AI Labb is the Data & AI division of TheAppLabb, a technology innovation firm that has launched 750+ applications over the past 18 years for global brands. We bridge the gap between AI's promise and business reality, delivering pragmatic AI solutions that create measurable business outcomes. Our approach is business-first: we start with client objectives and work backward to ensure every AI initiative delivers measurable ROI. We are looking for builders who share this philosophy. Our work increasingly takes us beyond individual AI use cases into designing enterprise agentic frameworks — the orchestration, observation, and decision-modelling layers that unify our clients' existing AI platforms into coherent, governable, brand-safe systems. We are connected into the Canada AI Alliance and Global AI Leaders Alliance networks, giving us cross-industry visibility into how the most demanding AI organizations are building.

Requirements

Must hold at least one of: AWS Certified Solutions Architect (Associate or Professional), Google Cloud Professional Cloud Architect / Professional Machine Learning Engineer, or Microsoft Azure Solutions Architect Expert.
Working knowledge of the other two clouds is required even if certifications are only on one.
Strong proficiency in Python. Must be comfortable writing production-grade code.
Strong foundation in traditional cloud architecture including networking, IAM, identity, serverless patterns, and infrastructure-as-code.
Proven experience with at least one hyperscaler AI platform (Amazon Bedrock + AgentCore, Vertex AI Agent Builder + ADK, or Azure AI Foundry) and operational fluency across all three.
Hands-on with vector databases such as Pinecone, OpenSearch, pgvector, or equivalent.
Demonstrated ability to build agents that utilize tools and function calling, with memory, planning, and multi-step reasoning.
Deep, current fluency with at least two of: LangGraph, CrewAI, Microsoft AutoGen, Semantic Kernel, Google ADK, AWS Strands, OpenAI Agents SDK, LlamaIndex Agent Workflows.
Practical experience with LLM observability tooling such as Openlayer, LangSmith, Arize, Langfuse, or Galileo.
Familiarity with policy-as-code and AI governance tooling — OPA, Cedar, NeMo Guardrails, Guardrails AI, Credo AI, or equivalent.
Working knowledge of MCP (Model Context Protocol) and A2A (Agent2Agent) and how they fit into enterprise agent architectures.
Ability to communicate technical concepts to business stakeholders, sit with a client's CTO and sketch an architecture, then sit with their CEO and explain why it matters.

Nice To Haves

High-level understanding of Snowflake, Databricks, Redshift, BigQuery, Synapse, and modern lakehouse patterns.
Experience with CI/CD pipelines and infrastructure-as-code using Terraform, CDK, or Pulumi.
Experience building production RAG systems with enterprise document collections, including evaluation and ongoing quality monitoring.
Familiarity with MLflow, Kubeflow, Vertex AI Pipelines, Azure ML, or Weights & Biases for model lifecycle management.
Prior work in financial services, healthcare, luxury retail, or government.
Prior experience at a Big 4, top-tier strategy firm, or specialist AI consultancy.
Published work, open-source contributions, GitHub repos, or speaking history in the agentic AI space.

Responsibilities

Design and code AI agents using cloud-native services (AWS Bedrock, Azure AI Foundry, Google Vertex AI / Gemini Enterprise Agent Platform) and modern agentic frameworks (LangChain, LangGraph, CrewAI, Microsoft AutoGen, Semantic Kernel, Google ADK, AWS Strands, OpenAI Agents SDK).
Implement reasoning, planning, and memory modules; configure LLMs to interact with external APIs, databases, and enterprise software to execute real-world tasks.
Design scalable infrastructure across AWS, GCP, and Azure, recommending the right one for each client's situation.
Select and optimize foundation models across providers (Bedrock, Vertex Model Garden, Azure OpenAI) based on cost, latency, performance, and data residency requirements.
Architect cross-platform frameworks for agent orchestration, observation, and decision modeling.
Design how agents communicate, hand off, and share memory across platforms, deciding when to use native platform orchestration versus a thin custom layer.
Architect tracing, evaluation, hallucination monitoring, drift detection, and cost telemetry across multi-platform agent workflows.
Design the policy catalog, guardrail enforcement model, and "where does this capability get built" routing rubric.
Participate in AI Discovery engagements to identify high-value opportunities within client organizations.
Translate business requirements into technical architectures that align with our outcome-driven methodology.
Lead the technical track of discovery engagements — stakeholder interviews, capability audits, framework design workshops, and executive readouts — and translate the architecture into clear business value.
Lead build-vs-buy analysis for each layer per client, with clear pros/cons grounded in their existing stack and team skills.
Maintain AI Labb's point of view on the orchestration, observation, and governance tooling landscape, refreshed quarterly.
Embed policy-as-code, identity, permissioning, PII handling, prompt-injection defense, and audit/observability into every architecture from day one.
Translate regulatory frameworks (EU AI Act, NIST AI RMF, ISO 42001, AIDA, Quebec Law 25, HIPAA where relevant) into enforceable controls within the framework.
Lead technical interviewing, selection, and onboarding for new hires within the AI Labb workstream.
Define technical standards and coding guidelines for our growing AI/ML team.
Mentor engineers and junior architects.
Contribute to knowledge transfer initiatives — building client capabilities rather than dependencies.
Integrate AI services into existing enterprise workflows and data pipelines.
Work fluently across enterprise platforms like Salesforce Agentforce, Snowflake Cortex, Microsoft Copilot Studio, ServiceNow AI Agent Studio, and the data infrastructure underneath them (Snowflake, Databricks, Redshift, BigQuery).