Database upgrades and migrations, Data Lake Architecture and Deployment, Machine Learning environment setup and ongoing management; expertise in AWS, Microsoft Azure, Cisco, Diamanti, Kubernetes, Kubeflow and more.
Identifying target data sources; ingesting, cleansing and joining data in batches or real time. Developing custom collection mechanisms to move data from sensors, machines, and other sources to a central repository for visualization and analysis.
Data Visualization, Analysis and Application
Visualizing data from multiple sources in common or customized methods; developing functionality for custom applications, use cases, reporting or alerting.
Data Modeling, Machine Learning and AI Prototyping
Model Development (Supervised, Unsupervised, Semi-Supervised), Testing and Deployment. Integration of model into existing software or custom UI/UX development; pilot ML/AI capabilities in limited/targeted capacity.
Machine Learning Model Production Deployment and Management
Deployment and ongoing management (tuning, retraining, security audits) of models deployed to production as well as supporting application connections (API) and functionality (UI/UX).
DIVERSE EXPERIENCE, PROVEN RESULTS.
Our perspective is informed by experience and success in multiple industries.
Reducing Unplanned Downtime
According the International Society of Automation, a typical factory loses between 5% and 20% of its manufacturing capacity due to downtime. Delta Bravo leverages data and context to generate more accurate predictions of the lifespan for a component given environmental conditions. When specific failure signals are observed, or component aging criteria is projected, Delta Bravo forecasts this to manufacturing operations, giving them several weeks advanced notice of when components should be replaced.
Delta Bravo is currently working with a global manufacturer with over 4,000 machines in 250 client sites worldwide. We’re improving Overall Equipment Efficiency (OEE) metrics. We’re helping the manufacturer forecast maintenance events and proactively train client operators to ensure higher uptime and longer machine life.
Reducing Cost of Poor Quality (COPQ)
Combining real-time monitoring and machine learning is optimizing process design, providing insights into machine-level loads and production schedule performance. Knowing in real-time how each machine’s load level impacts overall production schedule performance leads to better decisions managing each production run. Optimizing the best possible set of machines for a given production run is now possible using machine learning algorithms.
Delta Bravo recently completed an engagement that traced the root cause of component waste to a particular machine process. A machine learning model was applied to this process, enabling the manufacturer to predict if the process would have an 85% or better chance of success. If lower than 85%, the process was terminated and restarted. This reduced scrap rates, saving the manufacturer tens of thousands of dollars per month in wasted materials and improved time-to-production metrics.
Maximizing Throughput with AI-Driven Scheduling
Manufacturers that make multiple products on the same line often experience configuration-related downtime when processes are adjusted. Delta Bravo collects data from production line machines and previous order history to model a production schedule for incoming orders that maximizes throughput.
Delta Bravo completed an engagement with a plastics company responsible for over 400 products, ranging from tapes to silicone gaskets. Our model creates a production schedule for one week in advance and has shown a 20% reduction in downtime created by configuration changes.
Improving Accuracy of Demand Forecasting
Manufacturers that depend on sales teams and spreadsheets for Demand Forecasting often find themselves holding on to far too much inventory. Delta Bravo combines order history data with production data, logistics and even weather data (when appropriate) to improve the accuracy of Demand Forecasting models.
Delta Bravo worked with a market leader in refrigerated beverages to improve demand forecasting. In this scenario, the product went to a large retailer’s distribution center before being shipped to one of over 100 stores. The product has a freshness guarantee of less than 25 days and in some cases, various delays would impact the amount of “fresh time” on the store’s shelves, sometimes resulting in discounting that damaged the brand. Delta Bravo’s models helped sales people identify stores and distribution patterns that helped get product to stores faster and extend the “fresh time” of this delicious product.
Detecting Transaction Anomalies
Delta Bravo AI empowers administrators to be proactive if transactions start to spike above or fall below the norm so they can take action before an outage in service or a fraud scenario. Delta Bravo AI evaluates expected data volumes based on historical patterns and applies boundaries based on volume variation. This system is then used to compare real-time transaction value to expected volume, identifying anomalies in activity, resource consumption and more.
Credit Risk Analysis
Delta Bravo AI identifies credit payment and churn risks, leveraging various data sources to score applicants. Scoring and risk analysis is based on a number of factors like current income, employment opportunity, recent credit history, and ability to earn in addition to older credit history. Patters are also able to detect credit card “churners,” applicants who are rarely profitable for the card issuer. Delta Bravo AI also provides reason codes for credit decisions that explain the key factors in credit decisions, meeting compliance requirements for GLB.
Detecting and preventing fraud is a huge challenge for banks given the large variety of fraud types and the volume of transactions that need to be reviewed and manual or rules-based systems can’t keep up. Delta Bravo AI analyzes transactions and looks for indicators of suspicious behavior including transactions with dubious jurisdictions, suspicious companies or known parties. If fraud patterns are detected, Delta Bravo AI triggers processes to reject transactions outright or flag transactions for investigation and can even score the likelihood of fraud, so investigators can prioritize their work on the most promising cases.
Do you Know the Tipping Point for an Oncoming Emergency? Delta Bravo’s ability to understand photos and videos greatly helps in processing the mountains of data from surveillance systems or for “pattern-of-life” surveillance. Typical elements, activities and actions are modeled and analyzed so any unusual behavior or anomalies ranging from movements and gatherings to lighting conditions are detected and alerted on. This capability can be applied to high-transit locations or targeted areas for civic event coverage and more.
Optimize Emergency Response and Evacuation
Delta Bravo makes public infrastructure dynamically responsive to emergency scenarios. We are working with a city right now, leveraging forecasted flood plain patterns to adjust traffic signals for efficient evacuation of civilians and faster entry for emergency response teams. Delta Bravo can also compare and optimize human developed courses of action (COAs) alongside computer generated courses of action including the criteria for suitability, feasibility, acceptability, uniqueness and completeness.
Use Current Datasets to Improve Public Safety
Municipalities are already collecting massive datasets designed to help police, first responders and city planners. But are those datasets working together to protect more people and improve citizen quality of life? Integrating datasets, detecting anomalies and alerting the right responder prevents crime and saves lives.
Delta Bravo is working with a municipality today to combine License Plate Recognition data with Known Offender data to alert area officers when the wrong people are in the wrong places. Recently, this system alerted local officers of the presence of a known child predator in a school zone. It was also used to alert patrol officers of the presence of gang members in hostile territory. Officers arrived on the scene to defuse a potentially violent incident before it happened.
Improved Traffic Flow and Parking Management
Delta Bravo can ingest data from road sensors and traffic lights to built models for Adaptive Signal Control designed to reduce transit times and improve safety. This capability is estimated to reduce travel time by 25%, decrease vehicle stops by 25% and wait time by 40%. It is also being used to improve transit time for First Responders.
Delta Bravo can also be used to process license plate recognition (LPR) data, alerting parking management teams to current or future violations. This technology can also be used to forecast when available parking zones will be filled during civic events for better routing of traffic to available parking.
Leverage Sensor and Machine Data for Better Management of Water, Electric Resources
Delta Bravo processes data from disparate sources, such as pumps and filtration systems, through its algorithms to come up with optimal management and control protocols. This enables water managers to conserve energy, lower operating costs, and squeeze more life from water plant assets. AI can also be used to automate the inspection of sewer systems or the accuracy of meters. The same principles can be applied to municipality-run electrical grids and utility systems.