Forest Valley
About Forest Valley
We specialize in the development of complicated projects related to predictive analytics, Big Data, neural networks and machine vision.
Our company has a large team of qualified specialists from junior to senior level. We have a strong team of skilled developers: Java, Scala, PHP, Python, Ruby, Front-end (React, Vue); designer UI/UX, QA, analytics.
We specialize in the development of complicated projects related to predictive analytics, Big Data, neural networks and machine vision.
Our company has a large team of qualified specialists from junior to senior level. We have a strong team of skilled developers: Java, Scala, PHP, Python, Ruby, Front-end (React, Vue); designer UI/UX, QA, analytics.
Key Insights
Services & Focus
Industry Focus
Client Focus
Portfolio
Selection of contractors
Challenge:
The logistics platform has a very high turnover of customers and service providers. There is a daily replenishment of counterparties on average of 3% of the available clientele. The security screening service is overwhelmed with the amount of applications.Due to high demand, the screening procedures were often relegated to simple reviews of applicants’ forms. Meanwhile, it was still necessary to instantly select counterparties for each other.
Solution:
We had developed a special algorithm based on neural networks, for which our developers wrote a separate scoring module within the client software infrastructure. As a result, the service in real time recommends specific pairs of “customers” and “providers” with the highest probability to fulfill the order with the maximum composite KPI (time of submission, deadlines, punctuality, payment terms, disputes, customer evaluation). The average values ??of the composite KPI of completed orders increased 1.7 times.
Who is it beneficial for?
Logistics companies, banks, MFOs, UBER-like services.
Inventory assessment
Challenge:
The hardware chain store completely lacked the ability to track their inventories. There was no reconciliation between the amount of stock at warehouses and the product available at the store shelves. The main reason was the lack of inventory monitoring at the store level. It was further exacerbated by wholesale buyers, who could easily acquire an entire store stock, or customers relocating products between the shelves. As a result, the shelves often remained empty until the goods manager has passed by.
Solution:
Our solution was to use video cameras and machine vision system to continuously evaluate the amount of inventory left on the hands, and automatically fill out requisition forms to warehouses for stock replenishment. AI algorithms using rapid data analysis of purchases started making forecasts for demand of each item in the inventory, helping the management to make informative decisions for future purchases.
Who is it beneficial for?
Warehouse complexes, warehouses with direct access to retail customers, large retailers.
Laboratory examinations
Challenge:
Medical laboratory uses a special analytical technique for early detection and diagnosis of diseases. They screen human biological fluids for the presence of certain chemical structures (bio-markers), which can indicate various pathologies. Their other specialty is to diagnose skeletal diseases, e.g. osteoporosis, which reduces bone strength. X-ray is used for these diagnoses. The workload for specialists checking hundreds of images a day is colossal and due to natural fatigue, detection accuracy has decreased.
Solution:
We have created a system based on machine vision, which has a database of all types of disease markers and classifiers. It is capable of processing hundreds of images per hour. As a result, diagnostic accuracy has increased significantly. In case of uncertainties, images are reviewed by the doctors, otherwise alleviating their workloads.
Who is it beneficial for?
Medical institutions, manufacturers of medical laboratory equipment.
Product marking
Product marking
Challenge:
A large furniture manufacturer with an extensive dealer distribution network lacked operational information about the movement of finished products from the conveyor to the warehouse. The situation was exacerbated by the fact that the packaging alone did not allow exact identification, and product marking was only possible with an aid of the operator. When the production was at near or full capacity, the percentage of operators’ error became unacceptable.
Solution:
We developed and installed a machine vision system for product recognition. Integration with the 1C base (database) made identification of each item possible in real time. This allowed instant linkage of a product to a specific client and automatic marking directly on the conveyor.
Who is it beneficial for?
Conveyor (assembly) production lines.
Formation of product matrices
Challenge:
The regional grocery chain store is going to expand. Prior to opening a new store a main problem was: which of the dozens of product matrices, containing thousands of SKUs (stock keeping units), should be used?
Solution:
We developed a software application that uses neural networks to create a statistical model based on existing financial statements (e.g. income statement - sales) with incorporation of geographical and socio-environmental factors. The algorithm produces a rating system for all possible SKUs for the new store. The optimal matrix selection is then based on required number of SKUs that produces the highest rating.
Who is it beneficial for?
Any retail chains where stores differ in size (e.g. sq. ft.), social environment and assortment.
Forest Valley Reviews
Submit a review for Forest Valley
Location
78422777808