AI Software Developer - Intermediate

About Delta Intelligent Building Technologies (Canada) Inc. (formerly known as Delta Controls)

Delta Intelligent Building Technologies (Canada) Inc. (a subsidiary of Delta Electronics) is a leading building controls manufacturer with over 300 partners in more than 80 countries. For over three decades, we have provided innovative building automation solutions for commercial, healthcare, education, and leisure facilities. We delivered the world’s first fully integrated BACnet building solution for HVAC, lighting, and access control.

The Intermediate AI Software Developer is fully competent in software implementation, focusing on AI features and functionality for a project. They are expected to own and improve features in a project, with some supervision.

• Build and enhance Artificial Intelligence (AI) / Machine Learning (ML) features in projects, which may include Large Language Models (LLMs), Retrieval-Augmented Generation (RAG), and Agent orchestration tools like Strands.
• Develop backend services and pipelines in Python and other languages.
• Integrate cloud services, such as AWS Bedrock, into our AI features.
• Help design and maintain Agent knowledge bases, embedding workflows, and retrieval strategies.
• Provide feedback to the Technical Lead or Project Manager for development schedules.
• Review pull request of other team members, providing feedback when changes are needed.
• Contribute to the risk management activities of the Project Team.
• Contribute to code quality, testing, and documentation for the project.
• Participate in architectural discussions and propose improvements.

• Own end-to‑end delivery of mid-sized AI/ML features, from design to implementation to release, with minimal supervision.
• Break down requirements into technical tasks and milestones.
• Produce technical designs for AI services and agent workflows.
• Evaluate, select, and implement RAG patterns for specific use cases.
• Improve prompt, retrieval, and agent orchestration reliability through the use of guardrails, structured outputs, and tool‑use patterns.
• Mentor Junior developers through code reviews and technical guidance.

• To make implementation and design decisions within assigned feature area, selecting patterns and approaches that align with overall architecture.
• To review and approve pull requests, enforcing coding standards, test coverage expectations, and security considerations.
• To propose and implement improvements to AI workflows and supporting pipelines, within established architectural guardrails.
• To define and maintain technical tasks and estimates for assigned work.
• To recommend tooling and libraries to improve productivity, reliability, or quality.
• To make operational decisions for services they own, in alignment with team processes.

• Strong proficiency in Python, creating clean and maintainable code.
• Understanding of software architecture fundamentals and scalable design.
• Excellent debugging and troubleshooting skills.
• Experience with Git, pull requests, and resolving merge conflicts.
• Experience writing unit tests and working with frameworks like pytest.
• Ability to follow and contribute to coding standards and participate in code reviews.

• Interpersonal and mentorship skills.
• AI / Machine Learning
  
  Skills:
  
  Exposure to Large Language Models (LLMs) and working with model APIs, understanding of Retrieval-Augmented Generation (RAG) concepts, familiarity with embeddings and vector databases, experience with prompt engineering or building LLM-driven features, interest in multi‑agent systems or agentic workflows (Strands experience is a plus).
• AWS Cloud
  
  Experience:
  
  Exposure to AWS Lambda or serverless development, familiarity with AWS Bedrock or LLM‑related cloud services, experience with Dynamo
  
  DB and exposure to relational cloud databases (e.g., Amazon Aurora), basic understanding of Amazon S3, Cloud Watch, or IAM concepts, interest in working with AWS serverless and distributed architectures, experience with Terraform or Cloud Formation is a bonus.
• Nice‑to‑Have: RDF / semantic data or graph…