Our team of machine learning practitioners provide end-to-end solutions from data ingest to integrated operational solutions that run on customer premise systems. We have delivered solutions that run in the Google and Amazon clouds to ingest, process and automate machine learning.
We offer customers a specialized consulting services to develop a basic minimum viable machine learning model for their mission or a turnkey solution where the 4DV provides a complete solution based on using our licensed systems hosted in Google cloud in combination with the Customer Cloud or other commercial clouds. We offer custom on-premise solutions as well.
This service is to develop MVMS for a customer who needs help developing a feasible model and intends to transfer the learning and revise the model(s) to the next level on their own. This enables customer to fully protect their IP but leverage outside expertise.
Minimum Viable Machine Learning Model (MVM) is a machine learning model that is developed in the first stage of the development of machine learning models. The challenge in machine learning is to work with the data to get it into a useful form where a machine learning model can be applied. From our experience about 50% of the schedule is getting the data evaluated, filtered, error corrected and organized such that a machine learning model can be applied. After that a study can be done to determine a range of feasible applications and models to be applied. Sometimes the data lacks the inherent information to be useful and a model cannot be developed. If statistical data analysis performed on the corrected data indicates viable information content, then a MVM can be developed. This MVM demonstrations potential mission feasibility and is what we offer at this level of service. Often the MVM is feasible but the results are not good enough for the end mission application. For this service we help the customer to get positioned to go from feasibility to an integrated mission model.
Customer must provide and or identify the data to be used on the ML project and have completed a Google Discovery Workshop to identify viable missions/applications for their business. 4DV ANALYTICS will review the Cloud Discover Workshop report and customer proposed data sets to determine if the data set appears to be viable for the proposed application.
This service is to work with the customer to develop MVM learning models on customer supplied or open source data sets to demonstrate a minimum viable model or models. The scope involves data cleaning/formatting and organization, model development and training and providing output to demonstrate viability. The time period allotted is 3 months and can involve up to 6 FTEs.
Machine learning models do not always work with the data sets selected by the customer. Often the data is not adequate for a machine learning model. This may not be determinable until machine learning models have been developed to analyze the data. Data can be sparse, malformed, or not representative of a production environment. The false alarm rate can be too high or the data has errors that make it impossible to utilize the results. No model success can be guaranteed until the machine learning models have been developed, tested and proven to be robust against the data set. 4DV ANALYTICS attempts to mitigate model failure risk by an early look at the data prior to an engagement and review of the Google Discover Workshop results to determine model feasibility.
4DV ANALYTICS does a preliminary study on the customer’s data prior to the engagement to determine if mission feasibility is possible. Once 4DV Analytics has accepted the engagement we will work with the customer to get a MVM within a 3-month period.
We provide the development of MVM learning models. The customer brings the data sets and we develop the solutions using Google Cloud tools to process/correct/characterize and if necessary, embellish the data for machine learning use. The developed models are provided to the customer in their own Google Cloud Environments so that they can continue to develop their models into fully integrated mission models that include their proprietary application.
4DV Analytics is partnered with Google for this activity and leverages the relationship as a Google Partner to provide models to the customer
Google Cloud Team brings access to advanced technology and a knowledge base to insure that customers have access to latest information. We rely on Google to facilitate the Discover Workshop for the customer before we do the engagement.
4DV ANALYTICS team provides ML expertise and experience working big data and ML in the Google Cloud. Our team provides specialized and custom tools to support machine learning. 4DV ANALYTICS team works with the customer to help them develop their capability to upgrade the models to full mission capability and to use Google Cloud Tools. As a prerequisite the customer team needs to have basic software experience and a basic understanding of ML technologies.
We offer a fixed price, fixed LOE and schedule option. Our experience is that the MVM effort can be accomplished in a 3 month period, this aggressive schedule is dependent on the customer support. E.g., the customer needs to meet with the 4DV team on a weekly basis and provide deliverable data and feedback on a timely basis. The customer can purchase multiple models to increase the number of models for the mission application.
Up to 120 FTE days over a 3-month schedule from the start of the engagement. This work is done onsite or offsite at 4DVAnalytics discretion within a 3 month period. The MVM consulting engagement is contingent on prior completion of a Google Discovery Workshop and report which is provided to 4DVAnalytics prior to the engagement.
For this service we offer the use of 4DV ANALYTICS platforms to be used for customer mission applications. The 4DV platforms provide the customer a wide range of viable mission integration options using Amazon and Google Clouds and on-premise solutions. We offer customer specific machine learning capabilities using these platforms:
Varies with engagement. Online services are licensed on a per seat basis with a basic charge for setup and operational support. Cloud charges are applied on a cost basis using Amazon and Google clouds per current market prices.
Junior Analysts trained in strategic intelligence, business, marketing, history, psychology, statistics, economics, and cultural studies.
Senior Analysts in Executive and Senior Management positions, each with over 30 years of field experience in business, intelligence, and engineering projects