IT Intern

Your role at Tumi:  

The IT Intern will work with various teams across the IT department to deliver small, well-scoped projects supporting Samsonite Group’s Retail and E-commerce analytics. This position will work with digital engineers, data engineers and analysts to build pipelines in Digital Commerce products, Snowflake, model retail datasets using SQL, and use Python Plus AI-assisted development to accelerate delivery, testing, and documentation.

This person will also get hands-on opportunities to build and deploy AI agents to execute repeatable tasks at scale (with safe guardrails) and build lightweight ML models for forecasting and prediction (e.g., demand/sales forecasts, propensity indicators), starting in a sandbox and progressing to controlled production pilots.

 

• Retail Data Ingestion & Pipelines by building or enhancing ingestion for common retail feeds. Implement Snowflake loading patterns with optional automation (if applicable).


• Retail Data Modeling (SQL) by creating curated analytics-ready tables from raw feeds. Support omnichannel reporting by connecting various internal and external retail data elements


• Data Quality & Reconciliation (with Agent Automation) by implementing retail-relevant checks like uniqueness/deduping of customers, referential integrity (orders, product/store IDs) and anomaly detection (drops/spikes in sales, cancellations, returns). Build an agent to automate quality check and reporting. 


• Participate in system and integration testing in various other initiatives running in the departments


• Build and Deploy Agents to Execute at Scale - from learning to Production.
  
  
  
  • Phase 1 — Sandbox Learning (Weeks 1–3) Build a simple agent to generate validation SQL from templates and produce a report.
  
  
  • Phase 2 — Controlled Pilot (Weeks 4–7) Expand to scheduled runs, consistent output, and alerting. 
  
  
  • Phase 3 — Scale & Hardening (Weeks 8–12) Add guardrails (scope-limited actions), logging/audit trail, performance controls, and human review workflows. 
  
  


• Build ML Models for Forecasting & Predictions (Retail Use Cases) by working with a mentor to design and prototype a predictive model using curated datasets.

Typical Deliverables

• Feature set definition (lags, rolling windows, promos, seasonality) 


• Testing approach (time-based splits) 


• Model training and evaluation


• Forecast outputs written back into data tables for reporting 


• Basic explainability summary (top drivers / feature importance)

Deliverables & Success Criteria

By the end of the internship, the person should have delivered:

• Production-ready ingestion pipelines into Snowflake (raw → staging) 


• Curated retail data models (fact/dim tables) 


• A data quality suite with automated checks (SQL and Python) 


• At least one deployed agent that executes repeatable tasks at scale (logging and guardrails) 


• One forecasting/prediction prototype with documented accuracy


• Documentation: data dictionary plus pipeline runbook plus lineage plus model card 

A final demo showing measurable improvements (quality, timeliness, usability, or forecast accuracy)