GenBI Intern

Vistra Corp•Irving, TX

22h•Onsite

About The Position

The BI Retail team is building Retail AskBI Agent, a Generative BI platform that empowers business users to query enterprise data using natural language, automatically generating SQL, interactive visualizations, and executive-ready insights without needing to write code. The intern will contribute to the continued development and expansion of the AskBI platform, working on real-world business problems across Vistra's retail energy portfolio. This is a hands-on role that blends data engineering, AI/LLM integration, and retail energy business analysis.

Requirements

Expanded semantic layer covering at least 2 BI focus areas, generating new knowledge sources each validated through a suite of tested prompts to ensure reliable outputs
Document and execute a repeatable process to convert 3-5 manual email reports into Odin-generated automated distributions with scheduled email delivery
Snowflake Cortex AI Agent: a working native agent connected to the same Retail BI data sources as Odin, with a comparison analysis documenting accuracy, performance, and UX trade-offs
Final presentation to leadership showcasing project outcomes and platform impact

Nice To Haves

How Generative AI is applied to enterprise BI at scale
End-to-end data pipeline architecture (Snowflake → AI Agent → Visualization)
Prompt engineering best practices for structured data querying
Snowflake Cortex AI: building native AI agents within a cloud data platform
Business process automation: converting manual workflows into AI-driven pipelines
Platform benchmarking: evaluating multi-platform AI solutions to inform enterprise strategy
Retail energy business fundamentals: customer acquisition, margin performance, market dynamics

Responsibilities

Design and enrich Snowflake-based semantic schemas that power natural language querying across retail KPIs such as customer counts, sales forecasts, and margin performance.
Document business logic, metric definitions, and data lineage to ensure AI agents can accurately interpret and query the data.
Craft and refine system prompts and contextual instructions that guide LLMs to generate accurate SQL and business narratives.
Evaluate query accuracy across retail KPIs, identify failure patterns, and build a test harness of 50+ validated question-answer pairs.
Build reusable ECharts-based visualization templates that auto-render from queried datasets, including bar charts, line charts, and heatmaps.
Explore diagram generation for process documentation and design executive-ready chart formatting standards.
Build a native Snowflake Cortex AI agent that mirrors the AskBI capabilities currently deployed in Odin, leveraging the same underlying data sources and semantic models.
Benchmark accuracy, latency, and user experience between the Odin-based and Snowflake-native agents to inform platform strategy, and document a reusable deployment playbook.
Identify and catalog existing manual email-based reports across the Retail organization, mapping their content, recipients, frequency, and data sources.
Design and implement a repeatable process to replicate these reports as Odin-generated outputs with automated email distribution, and document an end-to-end migration playbook for future onboarding.
Participate in user feedback sessions to iterate on the conversational UX.
Create onboarding materials, use-case guides, and demo scripts to drive platform adoption across the Retail organization.