Hands-on AI-Architect (Snowflake Semantic Views)

The project focuses on building a robust and scalable architecture within the Snowflake AI Data Cloud. The primary objective is to enable business users to interact with complex transactional data using natural language. Key technical challenges involve scaling the existing semantic layer, defining the optimal structure and scope for Semantic Views, and developing a comprehensive data ontology that ensures high precision and trust in the AI-generated results.

Responsibilities

• Architect and Implement the Semantic Layer: Design and build scalable Snowflake Semantic Views that translate complex manufacturing and sales data into business-friendly concepts.
• Ontology Development: Create and maintain a detailed data ontology, including the definition of entities, relationships, and business-specific synonyms to improve model reasoning.
• Scaling & Performance: Resolve current challenges regarding the scope and structure of semantic models, ensuring they remain performant and return consistent results as the data volume and complexity grow.
• Verification & Governance: Establish processes for managing "verified queries" and metadata to ensure the AI assistant provides consistent and mathematically correct answers.

Qualifications and technical skills

• Expert-level Snowflake Knowledge: Deep hands-on experience with the Snowflake platform, specifically with its recent AI/ML features (Semantic Views, Cortex AI and Cortex Analyst).
• Semantic & Data Modeling: Proven track-record in building complex semantic layers, data ontologies, and advanced data models (e.g., Star Schema, Data Vault, or ODS).
• Manufacturing Domain Expertise: Experience working with large-scale manufacturing data (e.g., ERP, Supply Chain, or IoT telemetry) is highly preferred to understand the underlying transactional logic.
• Technical Hard Skills: Advanced SQL proficiency is mandatory.
• Soft Skills: Strong analytical mindset with the ability to bridge the gap between technical data structures and business requirements. You should be comfortable working in a collaborative, hybrid team environment.
• Problem-Solving: A "hands-on" approach to troubleshooting scaling issues and defining architectural best practices in a rapidly evolving AI landscape.

Summary of the practicalities

• Project start time: During Q1 2026
• Project duration: 4 months + possible extensions starting with a 2-3 week trial period
• Allocation: 100%
• Location requirements: Hybrid, with occasional face-to-face meetings in Helsinki
• Language requirements: English (required), Finnish (optional)

More information -> juho_witted // juho.sopanen@witted.com