Customer Story:

JLL's real estate ML services designed to scale with Qwak

JLL is a world leader in real estate services, powered by an entrepreneurial spirit.
JLL, build and invest in a variety of assets including industrial, commercial, retail, residential and hotel real estate. From tech startups to global firms, JLL's clients span industries including banking, energy, healthcare, law, life sciences, manufacturing and technology.
Qwak was brought aboard to enhance JLL's already impressive machine learning operation and enable delivery of complex models faster and more efficiently. The main requirement was to scale model delivery to production and enhance model accuracy without increasing engineering efforts.

10X
faster batch
inference
A/B
Model scenario
testing
↓$
Hundreds of thousands
$ savings yearly
About JLL
JLL is dedicated to transforming the real estate industry through technology-based innovation and develop and activate technology to make real estate work for the long-term benefit of JLL's people, clients and communities
Industry
Real Estate services
Proptech
Use Case:
Uncover real estate investment opportunities before they make it to market

Portfolio worth daily predictions via Automated valuation models (AVM)

Rent prediction valuation via deep learning text classification (Link to whitepaper)

Model Frameworks:
Xgboost
scikit-learn
Pytorch
Qwak stack:
Build system
Serving/Hosting
Analytics
Qwak's platform batch inference is significantly faster than other solutions we used in the past. Considering we have hundreds of these a day, the increase in speed allowed us to save hundreds of thousands of dollars"
Or Hiltch
VP Engineering @ JLL
The Challenge
Manual Model Training, Slow inference batch requests and scenario simulation
  • Model training and experiment tracking required a lot of manual work and alignment. JLL needed a central repository to efficiently track model experiments
  • Batch inference requests have to be run hundreds of times a day and each request took a long time which slowed down our development efforts 
  • The team developed new models that require online serving capabilities, JLL didn't have the infrastructure to support online models at scale
How Qwak Assisted?
  • Qwak Build allows JLL's team to centralize model experiment tracking and create a version repository to quickly compare between versions
  • Qwak's Inference The batch execution time was 10X faster in comparison to our previous solution. The ability to execute batch inference quickly saves both engineering time and money
  • Qwak Serving allows JLL to develop new types of models and deploy them as real time models at scale
Using Qwak allowed us to focus on creating value for customers and  rather than spending valuable time on our infrastructure setup.”
Or Hiltch
VP engineering @ JLL

Learn More About The Qwak Platform

 
Learn More