Enosis Outsourcing logo
Free Consultation
EndorsedendorsedOngoing

Media Mix Modeling Modernization for Scale AI

Services Covered:

Big Data Analytics and Business Intelligence • Custom Software Development
Project hero — replace with case study imagery
Industry
Industry

Advertising & Marketing

Duration
Durationinfo-circle

14 months

Budget
Budget

Confidential

Start Date
Start Date

1 July 2024

Team Size
Team Size

1-5

Engagement Model
Engagement Model

Offshore

Project Capability Score (PCS)

gauge meter
84/100
Strong

Project Capability Score (PCS) estimates how capable Armakuni is of successfully delivering a project like this, based on its past experience, track record, and ability to handle similar work. Learn more about PCS

Client

Scale AI

mapChicago, United States of America

Project Summary

Armakuni upgraded a media mix modeling platform for a marketing company by migrating architecture to DocumentDB and Amplify, rewriting the primary model and building two new models, and moving the front end to up-to-date React.js frameworks. They upgraded modeling components (PyMC3→PyMC5), simplified Docker files, integrated Airflow, and reduced the codebase by ~80%. The work enabled delivering insights 90% faster and helped the client double its client base.

Key Challenges

  • Upgrade the media mix modeling platform
  • Create new predictive models
  • Improve the overall application performance and architecture

Project Deliverables

  • Migration to Amazon DocumentDB and AWS Amplify
  • Rewritten primary model and two new predictive models
  • Migration to modern React.js frameworks
  • Simplified Dockerfiles and DAG/Airflow integration

Project Solution

Armakuni migrated the platform architecture from AWS Lambda, API Gateway, and S3 to Amazon DocumentDB and AWS Amplify, rewrote the existing model and built two new predictive models, and upgraded the front end to current React.js frameworks. They upgraded modeling libraries (PyMC3 to PyMC5), simplified Docker files, improved feature engineering, integrated DAG/Airflow workflows, and reduced the codebase by approximately 80%. The vendor provided a small dedicated team (five teammates) including data science and engineering support.

Project Outcome

  • Delivered insights 90% faster
  • Doubled the client's customer base
  • Reduced the codebase by approximately 80% and modernized frameworks
  • Improved client loyalty and decision-making using the updated models

Platforms

  • CloudCloud
  • WebWeb

Tech Stack

  • Amazon Web Services (AWS)Amazon Web Services (AWS)
  • React.jsReact.js
  • PythonPython
  • Apache AirflowApache Airflow
  • DockerDocker

Client Endorsement

Overall Review Rating

4.75star

Timeliness

Cost Rating

Willing to Refer

Quality of Deliverables

Our platform is better because of Armakuni. We now deliver insights 90% faster and have doubled our client base. They cut the codebase by 80% and moved us onto the latest React.js frameworks. Armakuni communicates well, and the quality of the work has improved.

Ed Sanchez

Director of Data & Analytics

Note: This endorsement is based on publicly available client feedback from external review sources.

More Projects by Armakuni

See All Projects by Armakuni