Our customer is a manufacturer of electronic devices and components for them, including mobile phones, TV remotes, DVD and CD players, digital cameras and others.

Detailed information about the client cannot be disclosed under the provisions of the NDA.

Challenge: Facing missed delivery deadlines and errors in strategic resource planning

Ensuring a well-functioning network of suppliers is crucial to guarantee timely order deliveries. Our customer has already optimized supply chain performance to maximize profitability by mitigating risks of fluctuating demand, inefficient operations, and volatile materials prices. Furthermore, they implemented rigorous planning and scheduling, comprehensive inventory control systems, and continuous monitoring for quality assurance.

Nonetheless, our client still faced missed delivery deadlines and errors in strategic resource planning. For improved accuracy and predictability in operational performance, they wanted an advanced DS & ML-based solution to collect and analyze large volumes of data and make realistic predictions on delivery terms.

Solution: DS and MLOps web extension to predict supplies and prevent delivery delays

As our client manufactures complex digital devices consisting of many parts (resistors, inductors, capacitors, transistors, diodes, etc.), they require stable, manageable supply chains with certain risks calculated. They wanted a broad overview of all previous interactions with partners, empowered by ML capabilities to digest and predict future shipments and prevent delays or interruptions in deliveries.

Based on that, Innowise suggested building a smart contract analysis platform that includes DS and MLOps to turn raw data into actionable insights. Our project team took full advantage of these technologies and implemented AI/ML in the supply chain to shield procurement processes and alleviate adverse effects.

Data pipelining

Once managers fill in all the information regarding certain partners (need for materials, delivery times, warehouse stocks, etc.), our platform produces predictions based on data pipelines. Thus, we implemented deep data analysis to catch data drifts and department divergence. Essentially, each step in the cycle creates an output that forms the basis for subsequent transformations, resulting in a continuous flow until each step is completed. Where appropriate, multiple processes are conducted in parallel to maximize efficiency.

Modeling layers

We developed a machine learning platform that estimates crucial factors affecting procurement process efficiency. Our team created a logical layer that clusters data into similar cohorts and trains models for each group. In addition, we incorporated an explainability layer to help an end user validate model behavior and better understand estimation.

Simply put, the solution’s flow can be described in the following way. Users input all the data regarding specific vendors like contracts’ IDs, required materials, order/delivery dates, current progress, and any auxiliary information. Then, based on the ML in supply chain algorithms, the platform analyzes indicated data and predicts procurement dates, considering the history of previous interactions, vendor reliability, and external risks. Predictive analytics, for example, might indicate when supplier stock levels are low or when delayed deliveries are likely to cause significant problems in the future.

Technologies

Back-end

TypeScript, Node.js, Nest, TypeORM

Front-end

TypeScript, React, Next.JS, MobX, MUI

Machine learning

Tensorflow, Keras, PyTorch, Scikit-learn, MLFlow

Data science

Pandas, Matplotlib, Plotly, Numpy

Databases

PostgreSQL

AQA, Manual

Process

At the first stage, our specialists clarified and redefined the customer’s objectives since the original proposal had many issues in terms of feasibility and end-use. Throughout the development process, our specialists applied additional AutoML approaches to boost model delivery rates. As our model received more samples similar to recent ones, we implemented a custom resampling technique that reduced the data drift effect.

Our project team worked on the Scrum methodology with bi-weekly sprints and everyday meetups. The project manager stayed in touch with the client, accommodating changes in the scope. All the tasks were tracked in Jira, with the PM assigning jobs and supervising overall performance.

Currently, the project is active, with our team working on improving output prediction and integrating ML supply chain modules.

Team

Project Manager

Business Analyst

Front-End Developers

Back-End Developers

ML Engineers

UI/UX Designer

QA Engineer

Results: Predictable ML supply chain modules with 630% lower risk of a production line halt

Innowise enriched the customer’s supply chain ML capabilities with a DSaaS extension to predict delivery terms. Thanks to ML and DS algorithms that consider the multitude of variables within a complex supply chain system, the client can now continuously monitor for potential procurement issues and plan shipments more thoroughly, preventing information silos. Thanks to the novel solution, the customer confidently manages supply chain processes without worrying about unforeseen complications or operational delays. Additionally, thanks to machine learning in the supply chain, our client can now make informed decisions that contribute to operational excellence and increased revenues across digital sales points.

Project duration

October 2022 - Ongoing

45%

reduction in disrupted deliveries

630%

lower risk of a production line halt

Related cases

Contact us!

Book a call or fill out the form below and we’ll get back to you once we’ve processed your request.

What happens next?

Having received and processed your request, we will get back to you shortly to detail your project needs and sign an NDA to ensure the confidentiality of information.

After examining requirements, our analysts and developers devise a project proposal with the scope of works, team size, time, and cost estimates.

We arrange a meeting with you to discuss the offer and come to an agreement.

We sign a contract and start working on your project as quickly as possible.

AI/ML in supply chain: 45% reduction in disrupted deliveries

Customer

Challenge: Facing missed delivery deadlines and errors in strategic resource planning

Solution: DS and MLOps web extension to predict supplies and prevent delivery delays

Technologies

Process

Team

Results: Predictable ML supply chain modules with 630% lower risk of a production line halt

Related cases

Contact us!

What happens next?

Let^●s develop
software together!

Make another request

Close

AI/ML in supply chain: 45% reduction in disrupted deliveries

Customer

Challenge: Facing missed delivery deadlines and errors in strategic resource planning

Solution: DS and MLOps web extension to predict supplies and prevent delivery delays

Technologies

Process

Team

Results: Predictable ML supply chain modules with 630% lower risk of a production line halt

Related cases

Contact us!

What happens next?

Let●s develop software together!

Subscribe to our newsletter

Make another request

Close

Let^●s develop
software together!

Subscribe
to our newsletter