endorsedOngoing

Real-Time Sensor Anomaly Detection Platform for Telecommunications

Services Covered:

AI and Machine Learning • Big Data Analytics and Business Intelligence • Data Engineering

Industry

Telecommunications

Duration

88 months

Budget

Confidential

Start Date

1 April 2019

Team Size

6-10 Employees

Engagement Model

Onshore

Project Capability Score (PCS)

78/100

Good

Project Capability Score (PCS) estimates how capable Dataeaze Systems 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

Confidential

India

Project Summary

Dataeaze Systems developed a real-time sensor anomaly-detection system for a telecommunications integrator, focusing on data ingestion, data management and AI/ML analysis. The engagement progressed through prototype/POC and iterative Agile releases and used technologies including HDFS, Kafka, Java and Python. The product was delivered within a reasonable time, with the vendor demonstrating flexibility, transparency and weekly reporting.

Key Challenges

Build a computer-vision-based anomaly detection system for sensor data
Ingest and process millions of data points per second at massive scale
Provide data ingestion, data management and AI/ML analysis to complete the product

Project Deliverables

Prototype/POC (iterative Agile releases 0.1 through 1.1)
Sensor anomaly-detection software
Data ingestion and data pipeline built on HDFS and Kafka
Machine learning models and model engineering (ML Engineer, Sr. Data Scientist)
Ongoing L1/L2 support post go-live

Project Solution

After a POC-based selection, Dataeaze delivered an iterative prototype-to-product build using Agile releases. They designed and implemented a data pipeline and ingestion layer (HDFS, Kafka) and developed machine learning models and software in Java and Python to perform near-real-time anomaly detection on sensor data. The team provided L1/L2 support and ran weekly reporting and Agile ceremonies throughout delivery.

Project Outcome

Product delivered in a reasonable amount of time
Many unknowns in the solution were identified and resolved by Dataeaze
Agile delivery with daily stand-ups and weekly reporting led to transparent project management
Client rated overall experience 8.5/10

Platforms

AI/ML Platform

Tech Stack

Apache Hadoop
Apache Kafka
Java
Python

Client Endorsement

Overall Review Rating

4.63

Timeliness

Cost Rating

Willing to Refer

Quality of Deliverables

“The product was delivered within a reasonable time. The Dataeaze team proved to be flexible when faced with challenges and project unknowns, coming up with solutions and new ideas. They reported on a weekly basis and used a transparent project management style.”
Shrirang Bapat
CTO
Date: 12/07/2020