endorsedCompleted

Machine-Learning Platform for Clinical Trial Design

Services Covered:

Custom Software Development • AI and Machine Learning • Natural Language Processing (NLP)

Industry

Information Technology

19 months

Budget

200K - 999K

Start Date

1 August 2018

Team Size

1-5

Engagement Model

Offshore

Project Capability Score (PCS)

74/100

Good

Project Capability Score (PCS) estimates how capable stermedia.ai 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

Trials.ai

San Diego, United States of America

Project Summary

Stermedia developed machine-learning solutions for an AI clinical-trial startup, building NLP pipelines and predictive models to analyze trial-related documents and predict trial outcomes. They delivered trained models and an easy-to-use framework that simplified deployment of analytical tasks. The vendor also supported web application development (Angular frontend, Python backend) and microservices deployment, and was praised for on-time delivery and strong communication.

Key Challenges

Revolutionize the manual clinical-trial design process by mining massive trial-related documents
Build models and tooling to predict trial outcomes from trial-design inputs

Project Deliverables

Trained machine-learning models for NLP tasks and trial-outcome prediction
NLP pipelines using transformer technologies (BERT) to process healthcare documents
Heuristics for identifying concepts in medical text
Web application frontend in Angular and backend services in Python
Microservices deployed in Docker and an easy-to-use analytical framework

Project Solution

Stermedia developed machine-learning solutions focused on two areas: natural language processing using BERT and other transformer models to process healthcare documents, and predictive models to forecast trial outcomes from trial-design inputs. Most development was in Python; Stermedia delivered trained models and heuristics, supported frontend (Angular) and backend (Python) software development, rewrote parts of code, built new functionality and provided microservices deployed via Docker. The vendor engaged a team that included two data scientists plus a scrum team to integrate ML deliverables into the web application.

Project Outcome

The vendor delivered an easy-to-use framework that simplified deployment of new analytical tasks
The team prepared heuristics for identifying medical-text concepts
The work was completed during a cooperation lasting almost a year and met timeliness and communication expectations

Platforms

Tech Stack

Python
Angular
Docker

Client Endorsement

Overall Review Rating

Timeliness

Cost Rating

Willing to Refer

Quality of Deliverables

“Not only did Stermedia deliver on time, but they also prepared an easy-to-use framework that simplified the deployment of the client's new analytical tasks. They provided a tremendous amount of value to the partnership through their flexibility and exceptional communication and collaborative skills.”
Tom Walpole
CTO
Date: 01/04/2022