Principal or Senior AI/ML Architect

Stratus

6d•Remote

About The Position

The Principal AI/ML Architect owns the architectural direction for the Intelligence and Agentic Workflows pillars of the Stratus Intelligence Platform. This role sets the technical direction for all AI/ML and agentic projects, ensuring the long-term coherence, scalability, and trustworthiness of the AI-enabled platform. It is a player-coach role, with significant time spent on hands-on design, prototyping, leveraging AI tools, writing reference implementations, and reviewing code. The core principle is Compounding Intelligence, where each AI and agentic project enhances the platform's defensibility. The role collaborates closely with the Senior Data Architect, AI/ML Engineering Lead, and other pillar architects. Key priorities for the first year include evolving the AI/ML platform (Intelligence and Agentic layers), leveraging AI in the development and delivery process to accelerate the roadmap, and ensuring foundational stability for production AI workloads.

Requirements

10+ years of software engineering experience.
5+ years in Principal, Staff+, or Architect roles at a B2B SaaS or AI/ML platform company.
Demonstrated track record of leading AI/ML and agentic architecture across multiple teams or product lines.
Experience designing systems where AI agents are first-class consumers (e.g., MCP servers, agent-facing APIs, retrieval-grounded interfaces).
Deep proficiency in Python and the modern AI/ML stack (LangChain, LangGraph, PyTorch, sklearn, vector stores, embedding pipelines).
Proficiency in C# / .NET and Azure for integration with Stratus's existing platform.
Hands-on experience with RAG, fine-tuning, prompt engineering, model training, and eval harness design.
Hands-on experience with agent frameworks (LangGraph, AutoGen, CrewAI, Semantic Kernel, or equivalent) in production.
Experience driving reliability and operational excellence for AI workloads in collaboration with a platform team.
Comfortable contributing to and reviewing code across the stack.
Excellent written and verbal communication.
Track record of mentoring senior engineers and elevating the technical bar.

Nice To Haves

Background in security architecture and threat modeling for AI systems (prompt injection, data exfiltration, agent abuse) and B2B SaaS.
Experience with MEP, BIM, construction tech, or other CAD/engineering domains.
Experience modernizing legacy ASP.NET MVC applications.
Experience with Autodesk products (AutoCAD, Revit, Fabrication, BIM 360 Docs, Forge) or other MEP/AEC/CAD software.
Domain knowledge of MEP, BIM, or construction technology workflows.
Prior experience in a Series B / growth-stage company navigating the transition from product-market fit to scale.

Responsibilities

Own the architectural vision and technical roadmap for Intelligence (data, learning, reasoning) and Agentic interfaces, workflows.
Design and evolve agent-based workflows, enabling reasoning-capable, adaptive, multi-step systems with strong trust patterns (grounding, evals, confidence scoring).
Develop and drive evaluation frameworks measuring model quality, agent reliability, drift, and platform effectiveness. Make AI workloads observable to engineering and product.
Architect for simplifying the core and customizing at the edge, with AI/ML capabilities being opinionated at the center and foundations for marketplace or platform-exchange patterns at the edge.
Architect agentic UX and enable just-in-time interfaces where the platform generates tailored views, avoiding over-investment in heavy custom UI work.
Build evaluation frameworks: model quality, agent reliability, drift detection, eval harnesses, observability for AI workloads.
Set engineering standards through working reference implementations, in-repo templates, and patterns for AI/ML teams.
Lead the polyglot AI stack: vector search, embeddings pipelines, RAG architectures, agent frameworks, model lifecycle, MLOps.
Drive AI-augmented engineering practice and set the standard.
Engage directly with customer-facing teams and customers to ground AI architectural decisions in real workflows.
Mentor engineers and raise the technical bar through code review, design review, and technical coaching.