FIT Next-Sizing Scanner
We are developing the FIT Next-Sizing Scanner, a deployed tool in the Google Cloud Platform (GCP) that connects to BigQuery data and focuses on the capture of Next-Sizing events that occur to help planners better understand true historical demand. The solution leverages a phased, agile methodology for rapid development, testing, and delivery of a prototype for direct deployment onto DLA's GCP environment with Impact Level 5 (IL5) compliance. The solution will be a deployed user interface frontend and backend both powered by cloud run that interact with underlying data stored in Bigquery.
Defense Logistics Agency (DLA)
AI Applications
The challenge our customer brought us.
The Defense Logistics Agency (DLA) faces inaccurate apparel sizing forecasts for new military recruits due to recent upward trends in recruit sizes. Current forecasting methods lead to invalid pre-positioned stock, forcing DLA to fulfill demands through special orders with significantly longer lead times (90-120 days). This impacts service readiness and supply chain efficiency. The DLA also has limited communication with recruit centers when it comes to data sharing, which adds complexity when trying to manage the combination of current customer needs and future expected demand.
Next-Sizing is when a recruit receives a uniform that is improperly sized because the optimal size for them is not available. It is particularly troubling because it is somewhat frequent due to the vast number of SKUs supported and it causes a self-propagating distortion signal in future forecasts. The problem being that the demand for the item that ran out of stock is only as high as what was available, whereas the “true” demand for the thing is in actuality higher, but a simple forecast will use the erroneous number and continue to predict a number that is too small for one SKU, and a number too large for others. This causes the issue to repeat itself.
What we designed and delivered.
A rapid AI solution to create a data-driven, repeatable, and accurate process for predicting apparel size distribution.
The difference it made.
By capturing and quantifying these signals, FIT helps improve the underlying forecasting tools to more accurately align with customer needs and allow the DLA to more effectively prepare the appropriate stock.
18
Other programs we’ve delivered.
Next Generation Threat System (NGTS) GPT
Learn More
1
Generative AI Logistics for Communication and Execution (GAILFORCE)
Learn More
2
Augmenting CILEMP to Enable Fleet Autonomy with Generative AI (CILEMP-FLAG)
Learn More
3
Goal Reasoning for Satellite Constellation Decision Making (GRANITE)
Learn More
4
Deep Reinforcement Learning
Learn More
5
AI-Powered Multi-Asset, Multi-Domain Operations Planning and Replanning Software (AMORE)
Learn More
6
Minefield Planning Evaluation and Gaming Suite
Learn More
7
Information Technology Services (LITeS) III
Learn More
8
Cyber Warning and Operations Center (CWOC) Research and Analysis
Learn More
9
Intelligence, Surveillance, and Reconnaissance Operations (ISRO)
Learn More
10
All Source Intelligence Cyberspace II
Learn More
11
Continuous Interactive Learners for Mission Planning (CILEMP)
Learn More
12
DARPA Explainable AI
Learn More
13
NIST Privacy Engineering Program
Learn More
14
U.S. Census Disclosure Avoidance System
Learn More
15
NAVAIR Next Generation Threat System (NGTS)
Learn More
16
Enterprise Master Management Agents (EMMA)
Learn More
17
Journal Vouchers (JV)
Learn More
19
Annual Financial Report (AFR)
Learn More
20
Translating All C To Rust (TRACTOR)
Learn More
21
Anticipatory and Adaptive Anti-Money Laundering (A3ML)
Learn More
22
Genomic Project
Learn More
23
Gemini for Government
Learn More
24
Logistics Information Technology Enterprise Solution (LITeS)
Learn More
25
Strategic Command Information Technology Lifecycle Support (SCITLS)
Learn More
26