The Role
Build and maintain Snowflake semantic views, Cortex Analyst and Agent configurations, and secure MCP integrations for AI-powered analytics. Own evaluation, accuracy, reliability, cost optimization, monitoring, and remediation for LLM-generated analytics. Develop Streamlit applications, integrate trusted internal and third-party data tools, and collaborate with data platform, GTM operations, and sales engineering teams. The role is onsite in Medellin and requires strong SQL, Python, Snowflake, dbt, Salesforce data model, and LLM analytics expertise.
Summary Generated by Built In
Senior AI Analytics Solutions Engineer
About the client
Our client is a leading global cybersecurity company. Its Digital Transformation Office is building the company's internal AI platform, with an AI-native way of working and security built into everything it ships.
About the role
You will make the client's AI analytics accurate, trusted and cheap to run. You will build and tune the semantic layer, conversational analytics and AI-powered go-to-market products that leaders use to find, grow and keep customers. You will report to the Senior Director of AI Engineering & Architecture, working with data platform, GTM operations and sales engineering teams: they own the business domain, you own the AI layer.
What you'll deliver in your first 90 days
• An evaluation harness for conversational analytics: golden question sets, accuracy scoring and regression checks on every semantic view change.
• Semantic view health monitoring covering accuracy, latency, credit cost and per-user attribution.
• A review and fix pass on existing semantic views and verified queries, such as fiscal-year filters, aggregation consistency and routing overlap between connectors.
What you'll do
• Design, build and maintain Snowflake semantic views and Cortex Analyst and Cortex Agent configurations that users query through Claude and MCP.
• Build and manage first-party MCPs that expose trusted internal data, tools and workflows through secure, well-documented interfaces.
• Evaluate, integrate and govern third-party MCPs, including the infrastructure for discovery, permissions, monitoring and lifecycle control.
• Own accuracy for AI-generated analytics: evaluation, backtesting, failure analysis and remediation.
• Improve the reliability, cost and architecture of existing AI analytics products, moving workloads from full generative output to analytics plus targeted AI insight where that is cheaper and better.
• Build lightweight AI analytics apps in Streamlit in Snowflake where they beat the alternatives.
• Instrument products for cost, performance and usability, and report on them.
What's out of scope
General BI reporting and dashboard support, data ingestion into Snowflake and data dictionary ownership, which sit with the data and analytics teams.
How the team works
• AI-native: Claude, Claude Code and Cowork are the default way of building.
• Security first: least privilege, auditability and change control apply to everything shipped internally.
• Working in the open: design notes in Confluence, work tracked in Jira and decisions written down.
Tech environment
Snowflake • Cortex Analyst / Agents / Search • Semantic views • dbt • SQL • Python • MCP • Claude • Streamlit • Salesforce data models
Requirements
Required
• 7+ years in analytics engineering, data engineering or ML engineering, with production ownership.
• Expert SQL and strong Python.
• Snowflake in production, including Cortex (Analyst, Agents or Search) and semantic views.
• dbt for modeling and testing.
• Built and shipped LLM-powered analytics or text-to-SQL, with an evaluation method you can explain.
• Comfortable with Salesforce opportunity and account data models.
• Advanced English.
Nice to have
• GTM analytics: propensity, expansion or churn models and how sellers act on them.
• Model evaluation methods: precision and recall over time horizons, uplift and holdout design.
• Streamlit, Tableau or other analytics front ends.
• RAG pipelines with Cortex Search or a vector database.
• Snowflake credit optimization and cost attribution.
Benefits
• Permanent contract (contrato a término indefinido) with all legal benefits
• Competitive salary
• Private health insurance (medicina prepagada)
• Work on the client's internal AI platform with an AI-native team
• On-site role in Medellín with parking
Skills Required
- 7+ years of experience in analytics engineering, data engineering, or ML engineering with production ownership
- Expert SQL skills
- Strong Python skills
- Production experience with Snowflake, including Cortex Analyst, Cortex Agents, or Cortex Search and semantic views
- Experience using dbt for modeling and testing
- Experience building and shipping LLM-powered analytics or text-to-SQL systems with an explainable evaluation method
- Familiarity with Salesforce opportunity and account data models
- Advanced English proficiency
- Experience with GTM analytics, including propensity, expansion, or churn models
- Knowledge of model evaluation methods such as precision, recall, uplift, and holdout design
- Experience with Streamlit, Tableau, or other analytics front ends
- Experience building RAG pipelines with Cortex Search or a vector database
- Experience with Snowflake credit optimization and cost attribution
Am I A Good Fit?
Get Personalized Job Insights.
Our AI-powered fit analysis compares your resume with a job listing so you know if your skills & experience align.
Success! Refresh the page to see how your skills align with this role.
The Company
What We Do
Roca Alliances is a consulting agency dedicated to helping global companies build geographically diverse teams and scale their operations by outsourcing or nearshoring to Latin America. With over 15 years of expertise, the company utilizes a proprietary approach to analyze, assess, and identify ideal locations across the Americas, fostering growth and scalability for clients in various industries including technology, media, and marketing.







