Lately Machine Learning (ML) technologies have evolved rapidly and are achieving an ever wider business adoption. ML algorithms are already applied routinely to problems, such as document classification and sentiment analysis, object detection in images, human pose estimation in videos, demand forecasting, chat bots and others. Quite often, the opportunity to apply ML to a business problem is not obvious and requires proper analysis and formulation in order to arrive at a successful solution. Data ingestion, cleansing, storage and analysis usually goes hand in hand with Machine Learning. We truly believe that we possess the right balance of academic knowledge and applied skills to deliver a solid ML business solution.
We can analyze your business and technology landscape and help you define the right approach to solve the problems at hand. Whether you want to gain competitive advantage, optimize costs or innovate to challenge the status quo, we know that every business case is unique and we want to hear about yours. On top of ML, Mopano's consulting portfolio includes the full spectrum of services, starting from design thinking and requirements specification, through project management, front-end and back-end development, up to documentation and trainings. Our team has broad experience in various domains, such as Telecom, Transportation, Banking, Automotive, Utility, Retail, Travel and Agribusiness and we are eager to learn even more.
Business critical systems often need to be extended and are usually not well integrated with each other. This can hinder innovation and put a heavy toll on the company's efficiency, making it hard to adapt to new technology and business challenges. Our team has a lot of experience in various enterprise integration technologies, inlcuding BMP, Workflow, Messaging and Process Automation. We can also help you take advantage of modern cloud technologies and paradigms, such as containers, microservices and serverless technology. Last but not least,we can also integrate a tailored ML solution into your software landscape, business processe and user interface, to achieve a complete, end-to-end solution.
We are a team of highly motivated professionals with broad experience in various industries. We fit well in both dynamic startup environments and established work processes at big companies. We welcome challenging projects that have the potential to disrupt the status quo.
Dmitry has been working in the field of Machine Learning for more than 10 years. He has his M.Sc. in Computational Neuroscience from TU Berlin and Ph.D. in Robotics from University of Stuttgart, Germany. Dmitry has created Data Warehouses for German Fintech companies and has extensive experience in data engineering, analysis and modeling. Recently he devoted himself to the field of Computer Vision, Deep Learning and Object Detection.
Dobromir is a software industry professional for more than 10 years, with solid experience in software development and management of projects and people. His background ranges from purely back-end solutions to full stack web development. Dobromir’s interests moved in the direction of AI and ML in recent years - general regression and classification problems as well as the application of ML in IoT. He has been a lecturer at Technical University of Sofia and several private software academies.
Svetlio is a senior software developer. He has worked on a variety of big enterprise projects for companies such as IBM, Lufthansa, British Gas and CERN. He holds a B.Sc. in Mathematics and Computer Science from Technical University of Sofia and M.Sc. in Information security from Sofia University. Svetlio has extensive experience in the Java ecosystem and has worked on several Big Data and IoT projects. He also has deep interest in applied Machine learning in the area of forex trading.
Ivo is a goal oriented professional software consultant, with more than 18 years experience in analysis, design, development and architecture of large-scale distributed software solutions. He helps companies turn big data and artificial intelligence projects into business success. He believes that machine learning will change companies in way that most of the domain experts can't even imagine today.
Antony is a solution architect with over 17 years of professional experience. He is currently focused on enterprise system integration at clients like Telecom Austria, Deutsche Bahn AG, Paul Hartman AG, etc. Over the years he has completed projects in many different industries including Banking, Transportation, Telecom, Retail and Utility services. He believes that customer thrust is hard to earn, so he never fails to deliver what has been promised regardless the effort needed.
Tina is a marketing specialist with more than 13 years of professional experience. She holds a Master's degree in International Economic Relations. Her career has left a positive impact on several big in international companies, such as OMV - an Austrian gas group and Kaufland - a big German retailer. We are now extremely happy to have Tina as our head of marketing and PR.
In 2018, one of our affiliates introduced us to the CEO of MOTORcheckUP. It is an interesting company that sells quick and easy testing kits for engine oils and other car fluids. The customer would just place a drop of oil on the test set and wait for it to mature. Then depending on the color, structure and shape of the spot they can assess the condition of the car engine. The company was already successful on the local market, but had trouble scaling up, due to the manual process of test evaluation. The CEO asked us if we can automate the process of test evaluation. He was discouraged, because several other companies had already tried and failed with this task. We accepted the challenge and time-boxed the first stage of the project to mitigate the risks. First we tried some existing cloud services, hoping to land at a quick PoC, but unfortunately none of them produced good results. Then we decided to explore the state-of-the-art in DeepLearning, possibly applicable to the task. Initially, we had some problems due to dirty input data. The devices that were used to capture the test photos needed some tuning, to improve the quality of the pictures. Then, after some additional data cleansing, one of the models was already showing promising results. We elaborated on that and finally came up with a solution based on Tensorflow that was already good enough for deployment to production. Our customer was very happy and asked us if we can build a full-stack solution, and integrate the Deep Learning model into it. As requested, we implemented all required front-end and back-end components and wired them with the Deep Learning model. Then we deployed the complete, containerized solution to the cloud preferred by the customer. The end result for our customer was that they were able to expand internationally, setting up subsidiaries in several countries around the globe (UAE, Turkey, Russia and others).
Back in 2016 we got an inquiry from the owner of W-Reservation - a young tour operator company, focused on the Eastern European market. They were growing fast and were facing a problem with the management of accounts receivable (AR). Their business model was simple. After a successful reservation, hotels should pay the commission within the contract term. However, payment were often delayed (DSO) and sometimes not paid at all (bad debt). W-Reservation wanted to address this issue in a flexible way that fits their business model. Namely, they wanted to balance the expected profits versus expected bad debt by adjusting the exposure of the hotel to their users. We created a machine learning model that would take into account factors, such as past bookings, payments etc and calculate an exposure score. We then integrated the model into the existing Java backend and added management UIs to facilitate the operations of the system by the backend users (e.g. manually adjust the score if needed). This resulted in a significant decrease of bad debt, while sustaining the sales volume. The net effect was increased profits without any negative impact on customer satisfaction and retention. After this successful project we maintained our relationship with W-Reservation and helped them with other challenges, such as integration with several GDS systems.
In the summer of 2014 we attended a DataScience event, where met the CEO of SmartShopper - a young startup company with the mission to transform the in-store shopping experience. Their idea was to aggregate product ratings from different sources and distill them into a unified product score. The CEO was concerned that many products lacked existing ratings at amazon.com, bestbuy.com etc, leading to inaccurate scoring and unhappy customers. On the other hand these same products were broadly discussed at different forums and articles. He wanted to use these additional sources to calculate ratings based on the sentiments for the respective product. However, this was an order of magnitude more complex than just using the official APIs of his usual sources. After careful considerations of scope and timelines, we drafted a project plan for a PoC that would fit the limited budget of his early stage startup. Our team tried several different approaches, ranging from classical NLP solutions to Deep Neural Networks managed and installed the one that worked best to SmartShopper. The sentiment analysis solution that we implemented was a success and played a crucial role in the future product growth.