Helping hiring managers make smarter, faster decisions with ML-powered candidate insights
Overview
To move closer to Walmart’s goal of hiring within 24 hours of an application, we designed Top K Recommendation — a tool that used ML-powered insights and a modern candidate comparison interface to help hiring managers make faster, smarter hiring decisions. The solution achieved 100% adoption across 4800+ stores and reduced time-to-hire by 50%.
Context
Walmart processes millions of job applications annually, with a vision to shorten the time-to-hire from weeks to within 24 hours.
Hiring managers needed better tools to evaluate large pools of candidates quickly and accurately.
Top K Recommendation was designed to provide clear candidate comparisons and leverage machine learning insights to improve decision-making speed and quality.
Problem statement
Store Managers, with their duties to keep a store running smoothly, are unable to spend the time required to hire a candidate that best fits the needs of the store. This adds significantly to the time taken to make a hiring decision and thus to hire.
How may we bring down the time to hire down to 24 hours?
Issue 1
Hiring managers struggled to process large candidate pools efficiently
![](data:image/svg+xml,<svg display="block" role="presentation" viewBox="0 0 24 34" xmlns="http://www.w3.org/2000/svg"><path d="M 19.729 19.735 L 9.864 29.605 L 0 19.735 M 9.864 0 L 9.864 28.615" fill="transparent" height="29.60500061035156px" id="gFf8LlmTP" stroke-dasharray="" stroke-linecap="round" stroke-linejoin="round" stroke-width="3.947" stroke="rgb(207, 190, 169)" transform="translate(2 2)" width="19.729000000000003px"/></svg>)
Time to select candidates dropped from 6 days → 1.5 days, and time to job offer improved from 8.87 days → 2.08 days
Issue 2
Candidate insights were hidden or fragmented across multiple tools
ML-powered Top K surfaced predictive signals directly in the UI, improving quality of hire from 41% → 50%
Issue 3
Requisition management was repetitive and inefficient
Hierarchy-level requisition creation reduced total requisitions by ~90%, delivering $550K in YoY savings and supporting scale across 4,800 stores
User Needs
I need a way to compare multiple candidates side by side
I want hidden insights about candidates surfaced automatically
I need a workflow that reduces repetitive requisition creation
Process Steps
Step 1
Research & Ideation
Conducted interviews with business stakeholders and store managers, refined workflows based on personas, and iterated through agile cycles
Step 2
Design & Iteration
Tested multiple low-fi mockups with hiring managers. Collected qualitative data through usability studies in pilot stores
Step 3
Collaboration with Machine Intelligence
Partnered with Walmart’s ML team to integrate candidate insights directly into the design, creating a feedback loop to improve the model
Step 4
Scaling
Piloted in select stores, then scaled to nationwide deployment after iterative updates based on feedback
Final Designs
Impact
100% adoption across 4800+ stores
Reduced time-to-hire by 50% (14+ days → 2 days)
Reduced requisitions by ~90% through hierarchy-level creation
Learnings
Machine learning insights are most effective when paired with clear, trustworthy UI design
Agile cycles and in-store testing drove rapid, meaningful iteration
Simplification at the workflow level (90% fewer requisitions) can have massive downstream impact
Learned how to design for scale: thousands of managers, millions of candidates