Our AI already works in production. Now we need the architect who makes it exceptional — and leads the team that gets it there.

AIAEC is an AI-native engineering technology company building agentic AI for AEC (architecture, engineering, construction). Our products reason over large volumes of complex technical and regulatory information and read architectural drawings — for professionals who need answers they can trust, cite and act on.

We're seeking a Lead / Principal AI Architect to take full technical ownership of AI at AIAEC and lead our entire AI/ML team: agents, RAG and GraphRAG, knowledge graphs, computer vision and deep learning, evaluation, and the AI infrastructure underneath them, on Azure and/or AWS. Our systems are in production — your job is to own them, raise their accuracy, reliability and scale, design what comes next, and lead the team that delivers it.

What makes this role different: in our domain, a confident wrong answer is worse than no answer. Correctness, traceability and evaluation are not nice-to-haves — they are the product. This is a hands-on role, not a diagrams-only one, and it is designed to grow into Head of AI.

What You'll Own

- AI/ML team leadership. Lead the entire AI/ML team — agents, retrieval, computer vision and ML engineers. Set priorities, review designs and code, mentor and drive engineers, and hire the people who scale the team.
- AI technical direction. The AI architecture and roadmap across agents, retrieval, knowledge graphs and computer vision — which models, techniques and vendors we bet on, and when to change course.
- Every agent, current and future. Design every current agent, and proactively propose new agents and AI capabilities that create value for customers. Own the agent architecture built on LangGraph: orchestration, state and memory, planning, tool use, multi-agent workflows, structured outputs, guardrails and human-in-the-loop patterns — designed for reliability, not demos.
- The whole RAG stack. Own retrieval end to end — vector, hybrid and GraphRAG: chunking strategies, embeddings, hybrid search, reranking, metadata filtering, query understanding, and grounded answers with verifiable citations.
- Knowledge graphs. Own the design of our knowledge graphs — ontology and schema, entity and relationship extraction, graph construction and maintenance — and how they combine with vector retrieval to power GraphRAG.
- Computer vision and deep learning. Own the technical direction for computer vision and neural network models that extract, check and compare information from drawings and documents — combining frontier vision-language models with custom-trained deep learning models.
- Scale. Architect AI systems for a high number of concurrent users: low latency, high throughput, horizontal scaling, async and streaming execution, queuing, caching,



rate limiting and graceful degradation under load.
- AI infrastructure. Set up and own the AI infrastructure on Azure and/or AWS: model gateway and routing, LLM and model serving, GPU workloads, vector and graph databases, prompt and model versioning, tracing and observability, fallbacks, and cost and latency budgets.
- Evaluation. The evaluation strategy and framework: golden datasets, offline and online evals, LLM-as-judge with human calibration, regression gates in CI, load and performance testing, and accuracy metrics leadership can trust. No change ships unmeasured.
- AI security and safety. Defenses against prompt injection, data leakage and tool misuse; access control for AI systems; hallucination detection; and responsible AI practices.
- The hard calls. Build vs. buy vs. fine-tune; frontier APIs vs. self-hosted models; what to fix, what to rebuild and what to keep — and where added complexity is worth it versus where a simpler approach wins.
- The engineering bar. Lead design reviews and write architecture decision records; set standards for code, testing, evaluation and documentation across the AI/ML team.
- Partnership. Work with data engineering on AI-ready data, and with product and engineering leadership to turn AI capabilities into features customers rely on.

All About You
- Experience: 10+ years in software or ML engineering, including 3+ years building and running LLM-based systems in production — across several companies, so you've seen these patterns work and fail in different organizations.
- Agents & LangGraph: You've designed and shipped production agentic systems used by real customers — strong hands-on LangGraph experience required — and can walk us through their architecture, failure modes and the metrics you moved after launch.
- RAG, Hybrid & GraphRAG: Deep, hands-on expertise across vector, hybrid and graph-based retrieval: embeddings, reranking, chunking and metadata design — with evidence you've measurably improved retrieval quality.
- Knowledge Graphs: Hands-on experience designing and operating knowledge graphs and graph databases (e.g., Neo4j or equivalent), including ontology design and entity and relationship extraction.
- Computer Vision & Deep Learning: Solid foundation in neural networks and computer vision — CNNs, vision transformers, detection and segmentation — with hands-on PyTorch experience training and deploying models,



plus vision-language models and document understanding.
- Scalability: You've architected and operated AI or backend systems serving a high number of concurrent users, and know how to keep latency, cost and reliability under control at scale.
- AI Infrastructure: You've set up and owned production AI infrastructure on Azure and/or AWS: model serving, GPU workloads, LLM observability, model gateways, caching and AI security.
- Evaluation: You've designed evaluation frameworks from scratch and used them to make shipping decisions — implemented across real teams and production systems, not just read about.
- Engineering Excellence: Expert Python; strong distributed system design; APIs, async programming, containers and Kubernetes, CI/CD and Terraform. You still write production code.
- Team Leadership: You've led an AI/ML team of 5+ engineers across different specialties, set technical direction, and made the people around you measurably better.
- Product Thinking: You spot where AI can create new value and turn those ideas into concrete proposals, prototypes and shipped agents.
- Battle Scars: You can walk us through an AI initiative that went wrong — a model, agent or retrieval system that failed in production — why it happened, and what you'd do differently now.
- Communication: You explain decisions, trade-offs and risks clearly to engineers and executives — in writing and in person.
- Education: Bachelor's degree in Computer Science, Engineering, Machine Learning or a related field, or equivalent practical experience.

Nice to Have
- AEC Domain: Experience with engineering drawings, CAD or construction documents.
- Regulated Data: Experience building AI over legal, regulatory or government information.
- Advanced Degree: Master's or PhD in Machine Learning, Computer Vision or a related field.
- Model Optimization: Fine-tuning, distillation, quantization and high-throughput inference serving.
- Startup Experience: You've helped scale an early-stage AI team and shipped fast without breaking customer trust.
- Public Work: Talks, writing or open source contributions in applied AI.

Why AIAEC
- Own AI at an AI-native company — your architecture is what every product decision stands on.
- A clear path to Head of AI: after 90 days, we intend to formally appoint you to lead AI at AIAEC.
- Work at the frontier of applied AI in an industry ripe for disruption — AEC technology is wide-open territory.
- Report directly to the CEO, with real authority over technical direction — not just execution.
- Flexible ways of working and a small, senior team.

If you hold yourself to a high technical bar and want to own AI that professionals rely on — let's talk. Pay: $80,300.00-$167,000.00 per year

Benefits

- Paid time off
- Work from home

Work Location: Remote

📌 LEAD / PRINCIPAL AI ARCHITECT (Montreal)
🏢 Sustainability Engineering Group
📍 Montreal