AI Developer(Google AI Must)

4+ Months Contract

Montreal, QC OR Mirabel, QC- 4 Days a week Onsite

Role Objective

The [North American Digital Division] is seeking a highly skilled, hands-on AI Developer to drive the development, coding, and industrialization of our Document to Diagram (D2D) solution and related Process Document processing workflows.

The successful candidate will execute the technical development lifecycle of this specific AI asset, moving an existing Proof of Concept (PoC) into a fully industrial, secure production environment.

The developer will be solely responsible for ensuring the D2D tool extracts process steps with high precision, maps them to BPMN compliant schemas, generates editable visual structures, and integrates seamlessly within the secure enterprise cloud ecosystem.

Note: No baseline training or upskilling will be provided on the core stack; the candidate must hit the ground running.

Key Activities & Responsibilities

Pipeline Integration & Coding: Implement the technical design of the Document to Diagram application, writing and deploying integration code that connects Vertex AI LLMs with frontend user interfaces, document repositories (Google Drive), and visualization engines.

Core Engine Development: Code and refine D2D parsing workflows using Python and GCP Vertex AI. Build and maintain the document ingestion logic to extract actors, activities, gateways, and sequences from unstructured PDFs and Word files.

Format Conversion: Author scalable, high-performance serialization scripts to map extracted process structures directly into structured, editable .drawio (XML-based) and Sparx Enterprise Architect (.xml) formats.

Agentic Logic & Conversational Refinement: Code agentic decision loops using LangGraph (or similar frameworks) to enable the tool's Conversational Refinement Mode. Implement advanced state management to retain conversation history and diagram contexts across chat sessions.

Interactive Canvas Design: Utilize knowledge of Canvas editions UI/UX to enable collaboration for

iterative diagram editing and updates.

Industrialization & Scaling: Implement production-grade pipeline capabilities including Bulk Processing Mode (Queuing, Deque, Pub/Sub), API scaling, automatic splitting logic for extra-long diagrams, and smart flagging for uninterpretable non-text source elements.

Secure Cloud Integration: Connect Vertex AI, user interfaces, and document systems securely adhering to deployment patterns such as Google IAP, OAuth2, and RBAC.

CI/CD, MLOps, & Testing: Set up unit tests, integration tests (specifically regression testing for diagram output accuracy), latency monitoring, and performance/cost tracking for Vertex AI endpoints.

Code Quality & Standards: Adhere strictly to clean code practices, maintain high test coverage,

participate in peer code reviews, and draft clear technical documentation for system APIs and developer handbooks.

Required Skills & Qualifications

Core Experience: Possess 3–5+ years of software experience with expertise in AI NLP Engineering and Document processing. Must have a proven history of developing and deploying production-ready RAG applications in enterprise environments, including document generation and Named Entity Reconciliation.

Cloud & AI Infrastructure: Mandatory, hands-on experience with Google Cloud Platform (GCP) and

Vertex AI, including Vertex APIs, prompt engineering, structured JSON outputs, and LLM orchestration.

Backend Development: Deep expertise in Python and modern backend frameworks such as FastAPI/ Flask, Pydantic, and schema validation.

Parsing & Schema Manipulation: Strong experience in custom document and diagram generation (XML/ JSON), structure manipulation, and parsing, specifically generating complex XML schemas compatible with Draw.io and Sparx systems.

AI Orchestration Frameworks: Hands-on experience building multi-step AI Agent systems using

frameworks like LangGraph and ADK to manage multi-step generation, evaluation loops, and state-history persistence.

Agentic Design: Proven AI integration experience maintaining clear separation between tools and skills within agentic development.

Enterprise Security: Familiarity with secure enterprise cloud integration, including Google Identity-Aware Proxy (IAP), MCP elicitation, and OAuth2.

CI/CD & Tooling: Expertise in agent coding using Claude code, Agy, and Gemini clit, along with

containerization and enterprise CI/CD toolchains like gcloud, terraform, docker, jenkins, and github.

Execution & Autonomy: Strong execution skills with a proven ability to work autonomously, deliver

features within Agile sprints, write high-quality code, and manage individual technical tasks efficiently.

Technical Communication: Ability to clearly explain technical workflows, code logic, and API designs to the AI Project Manager, Solution Architect, and Business Analysts.

Inputs & Project Dependencies

Company Project Management Methodology (Waterfall / Agile).

Internal Business Processes (specifically Program-Specific Process Modeling Standards).

Corporate AI Governance and best practices framework.

Existing D2D Proof of Concept (PoC) code and the Business Requirements Dossier (BRD).

Key Deliverables

A production-ready Document to Diagram (D2D) AI solution.

Comprehensive technical documentation and developer handbooks