AI Architect

ContactMonkey•Toronto, ON

17h•Hybrid

About The Position

ContactMonkey is seeking an AI Architect to join a dedicated AI pod focused on advancing the company's AI capabilities. The role involves setting technical direction, making foundational infrastructure decisions (vector store, evals, data warehouse, agent orchestration), and developing production systems. This is an architect role with hands-on coding responsibilities, focusing on shaping the AI trajectory of the company. The AI platform already includes features like Claude-powered template editing, a ConfidenceCheck service, Braintrust-backed evaluations, and feature gating. The next phase will involve grounded assistants, retrieval over customer data, richer agentic flows, and smarter data products for sending decisions.

Requirements

Credibility across AI, Cloud, Data, and Fullstack.
Experience building and operating LLM-backed features in production.
Deep familiarity with at least one major provider SDK (Anthropic, OpenAI) and comfort switching between them.
Understanding of prompt caching, structured outputs, tool use, and tradeoffs between agent loops and pipeline orchestration.
Experience running a real eval harness (Braintrust, promptfoo, in-house) against golden datasets and using it to block regressions.
Working knowledge of retrieval systems: embeddings, hybrid search, reranking.
AWS fluency at the architecture level (VPCs, IAM boundaries, secrets management, queue-based decoupling, cost control).
Comfortable with container-native deployment (ECS, EKS, or equivalent).
IaC experience (Terraform or CDK).
Experience operating production systems under real load and writing postmortems.
Ability to design a data pipeline end-to-end (ingestion, transformation, storage, serving), including batch and streaming.
Experience building data pipelines in at least one real organization.
Comfortable with SQL and non-SQL at a level that allows reasoning about query plans.
Experience tuning slow queries by analyzing query plans.
Understanding of when a data warehouse is necessary versus when read replicas are sufficient.
Proficiency in Ruby on Rails, Vue.js, Nuxt.js, Fastify/TypeScript.
Ability to write real TypeScript.
Understanding of API design beyond the happy path (versioning, auth modes, rate limiting, idempotency).

Nice To Haves

Fine-tuning or distillation experience.
Classical ML for ranking/classification.
Prompt-injection and jailbreak threat modelling.
Experience running workloads on GCP.
dbt, Airflow / Dagster / Prefect experience.
CDC patterns experience.
Multi-tenant data partitioning at scale experience.

Responsibilities

Set the technical direction for the AI pod.
Make foundational infrastructure bets, including vector store, evals, data warehouse, and agent orchestration.
Ship the first production AI systems.
Define AI platform direction, including model selection and routing, prompt and context architecture, evaluation strategy, cost and latency budgets, and safety/PII handling.
Design and implement retrieval and grounding infrastructure, including ingestion paths, chunking, embedding, and refresh strategies.
Develop agent and tool-use patterns, including multi-turn tool-calling loops, fallbacks, observability, and human-in-the-loop checkpoints.
Establish data engineering foundations, including data warehouse decisions, pipeline design for product analytics and model training/eval datasets, and ensuring data lineage.
Determine cloud and deployment architecture for AI services on AWS, including microservice patterns, GPU inference, and tenant spend isolation.

Benefits

100% employer-paid benefits + a Health Spending Account from day one
Work from anywhere in the world for up to 4 weeks
Stock option plan
RRSP Group Savings Plan
Generous vacation package
Personal development budget
One personal day + two volunteering days
Your Birthday off
Five health days per year
Beautiful downtown Toronto office for hybrid work—fully stocked with all the best snacks