Our Work
Anomaly Detection
Proactively discover and investigate anomalies and behavior, and payments.
Powered by data science, our anomaly detection methodologies are designed to improve compliance, lift integrity, and reduce error, waste, and fraud.
Monitoring and operations optimization
Situation
- With over 200,000 sensors streaming metrics from operational plants 24x7 into a data lake the platform was unable to cater for massive data volumes and support required data science query response times
- It was also supporting thousands of data science workloads against the sensor data (in support of platform performance, maintenance, etc.) which presented an operational risk
- Inability to onboard new datasets and use cases
- High operational and support cost
Solution
- New data lake solution architecture defined including benchmarking various solution options
- Implemented and migrated existing data to the new data lake with no impact to existing data science workloads and production
- Developed new access APIs to support ease of onboarding new use cases
- Implemented new security architecture
- Implemented new DevOps processes to improve speed and quality of software deployment
- Want to know more about our Managed Service Platform technologies click here
Benefits
Scalable, agile and flexible AWS data lake architecture able to support required performance requirements and ability to rapidly onboard new use cases
Improved speed and quality of software deployment
Dramatically reduced operational cost
Utility infrastructure monitoring and leakage detection
Situation
- A client with a very large water and waste network in New South Wales
- A widely distributed legacy SCADA network and were assessing the viability and benefits of rolling out a new geo-tagged IOT based low powered sensor network
- Required a scalable near-real time solution to quickly detect and locate water leakages and major burst detection in the pipe network which typically results in 10% of water loss in any year
- The client was looking for a technical partner who had the AWS cloud and data science / machine learning expertise to rapidly develop a proof of concept to identify water leakages
Solution
- An anomaly detection machine learning algorithm was developed which could identify a potential water leakage based on changes in pressure and flow in the pipe network
- Established the platform to ingest sensor readings, invoke the machine learning algorithm and present IOT sensor status on a map of the client's pipe network (based on AWS cloud services)
- New visualization tools allowed a user to easily analyze historical trends and view the network status over different time horizons. Healthy sensor status was shown as green and leakages identified as red over customer map overlays
- Want to know more about our Managed Service Platform technologies click here
Benefits
In a 6-week period, an end to end solution was developed from ingestion, machine learning algorithm development to visualization as a proof of concept
A PoC highlighted several issues with the existing sensor network in terms of time stamping of messages and network equipment which were critical findings for consideration to a larger scale product environment
Detecting and mitigating impact of fraud in messaging
Situation
- Lack of clarity on scale of fraud instances such as phishing and spam activity
- Regulators and Mobile Network Operators losing confidence in A2P providers
- Consumer trust declining
- Global Messaging volumes growing and forecast to be 3.5 trillion messages a year by 2023
- Related global losses circa $2b USD annually forecast to grow to $62b by 2023
- Clients sought greater understanding of customer behaviour and sentiment with proactive tools to manage compliance and customer experience improvements
Solution
- Deep analytics identifying fraud messages and senders
- Identifying anomalous messages in real-time
- Data enriched and streamed to live dashboard showing customer activity, message themes and irregularities
- Data store for additional data science analysis - predictive and prescriptive innovations
- Enabled a differentiated offering for alerting and fraud mitigation prior to customer impact
- API allowed for flags to be returned in message management
- Want to know more about our Managed Service Platform technologies click here
Benefits
High availability, redundancy, scalability, and resilient platform - inclusive of desired reporting features
Optimized cost profile for scale and performance targets
Modular serverless architecture proving zero lock in technology commitment for ongoing platform enhancement
Infrastructure supporting roadmap for innovative product development at scale
Critical IT infrastructure, payments, and communications ecosystem performance
Situation
- Near-real time monitoring and reporting across payment hubs, unified communications ecosystems and contact centres
- Required infrastructure to cope with speed and scale demands due to the quantity of data involved and analytic processing complexity
- Sought to modernize on-premise legacy platform to address new markets, improve speed-to-market and launch new product offerings
Solution
- Provides deep analytical capability with full SQL support as required
- Real time ingestion, storage, and provision for query engine - utilizing hot and cold data pathways to optimize performance and cost profile
- Data recency faster than 30 second target
- Scalable solution to accommodate high volume, peaking at 300,000 payment transactions per second during black Friday sales
- Want to know more about our Managed Service Platform technologies click here
Benefits
High availability, redundancy, and resilient platform - inclusive of desired reporting features
Optimised cost profile for scale and performance
Modular serverless architecture proving zero lock in technology commitment for ongoing platform enhancement
Infrastructure supporting roadmap for innovative product development at scale