Roman Panarin is a seasoned professional with extensive experience in computer vision, data science, and machine learning (ML). Roman has demonstrated a deep understanding of diverse AI domains with 6 years dedicated to honing his skills in computer vision, data analytics, recommendation systems, and predictive analytics, followed by 2 years focused on Natural Language Processing (NLP).
His career as ML Engineer started at Integra Sources, where he developed algorithms for fall detection, people detection, and player recognition, demonstrating his proficiency in real-time applications. Meanwhile, his tenure at CSort LLC involved creating classification algorithms and segmentation models for various industrial applications.
His academic pursuits are equally impressive, as he holds the position of Associate Professor at Altai State University. Roman is committed to shaping the next generation of machine learning professionals. He designs and delivers practice-oriented courses, ensuring students acquire the practical skills needed to excel in industry roles. Furthermore, his contributions extend beyond the classroom, with involvement in wildlife identification projects and automated remote sensing software development.
Currently serving as a ML Engineer at Mad Devs, Roman has led projects focusing on supervised text generation and pattern trading search systems. Notably, he developed a groundbreaking algorithm that generates summaries based on telephone conversations, achieving text quality indistinguishable from human-written content. Additionally, Roman devised a pattern trading search system capable of scanning vast cryptocurrency pairs, demonstrating his versatility across different applications. Also, he holds the esteemed position of Head of the ML Community at Mad Devs. Here, he spearheads initiatives to foster collaboration, knowledge-sharing, and innovation among fellow professionals in the field.
About
Education
Altai State University
Bachelor’s degree, Applied Mathematics and Computer Science
Master’s degree, Applied Mathematics and Computer Science. Master of biocybernetics, bioinformatics and software engineering
Industries
Data science & ML
HealthTech
EdTech
IoT
BYOD
Blockchain and crypto
Digital transformation
Articles and insights
Transfer Learning from Large Language Models (LLMs)
In a world where technology is advancing at an unprecedented pace, transfer learning stands out as one of the most promising and influential areas of machine learning. This approach to training models opens doors to faster and more efficient development, as well as flexibility in adapting to diverse tasks. This article will introduce you to transfer learning, technical nuances, applications across various business sectors, and future prospects. Discover how transfer learning is shaping the modern world of technology and the opportunities it unveils for us all.
ML
Digital Twins in Medicine
The Industrial Revolution 4.0 is growing in momentum with the rise of faster wireless technologies for data exchange, more data collection, and the development of advanced approaches to data processing. Thus, today we can get an incomparably greater quantity and higher quality of data about the world than ever before.This opens up fundamentally new possibilities and allows us to create new technologies that revolutionize how we approach the design, development, and application of new things in various fields, from manufacturing and transportation to medicine.These technologies include digital twins, which we will talk about today. In this article, you will learn:A digital twin is a virtual model of a real object, system, or process based on a variety of data collected from it by multiple connected sensors. We thus get a virtual model with all the real object's properties, characteristics, and states. And we can simulate various aspects of its operation with much less time, money, and energy.This approach is popular for a variety of reasons. From calculating the best usage patterns, resource allocation, and other important aspects for areas such as the transportation industry, etc. To pre-calculating critical aspects of operations, design flaws, and more in areas where the cost of error is unacceptably high such as aviation, medicine, and others.
ML
Boosting Hiring with AI Recruitment Software
Artificial intelligence (AI) is everywhere, from tools for individual use to large-scale enterprise solutions, many business owners are looking to see how they can utilize AI in their specific industries to improve their operations.For such business owners, staying up to date with the latest technology is crucial. Managers of smaller companies may wonder if these tools are only meant for large corporations. The answer is no. Businesses of any size can benefit from these existing tools to improve productivity, reduce costs, and streamline recruitment, which is the focus of this article.AI recruiting software offers many benefits to companies, including automating your workflow to engage, screen, and assess candidates more effectively. After examining this and other benefits, we'll share the most popular tools in AI recruiting available today.
ML
Everything You Need to Know About Edge Computing
Today, data is everything, so the best ways to collect, process and share it are critical. Edge computing in this context cannot be overestimated because it plays a key role in how we do it. Connected with other types of computing actively developing in parallel and often combined with it, we get improvements that make technologies develop faster and faster.Especially when you consider that edge computing widely supports and distributes many other advanced technologies, such as IoT and AI/ML, that are changing our world on an even more fundamental level. All this together gives us confidence that the world will change beyond recognition over the next decade.In this article, we will take a closer look at what edge computing is and what types of computing are developing along with it. How does edge computing work, what problems does it solve? And what technologies and entire industries are affected by edge computing? All this and more are in our article. Have a beneficial read!Edge computing is a way of handling data processing that revolves around working with information close to its source instead of depending on faraway data centers or cloud resources. It includes a variety of technologies and designs aimed at enhancing the efficiency of data processing by shortening the distance data travels before being processed.Let's delve into the unique characteristics that make edge computing powerful:1. Localized processingUnlike the cloud, edge computing processes data closer to its source, leveraging geographically distributed platforms. This is ideal for sensor networks (think pipeline monitoring).2. ProximityWith processing power located near the data source, users gain a valuable advantage: they can leverage network context information to make smarter decisions about how to utilize resources and services.3. Mobility supportWith the explosion of mobile devices, edge computing seamlessly communicates directly with them, enabling location-aware services like fog-based vehicle safety and edge-based disaster management.4. Low latencyEdge computing minimizes delays by keeping resources and services local. This is crucial for resource-intensive applications like connected cars, remote health monitoring, and industrial control systems.5. Heterogeneous ecosystems Edge computing embraces diverse technologies. Different platforms, architectures, computing methods, and communication protocols can coexist within edge computing elements – end devices, edge servers, and networks, e.g.6. Reduced bandwidth burdenIt tackles bandwidth-intensive tasks like video analytics closer to the source (surveillance cameras, e.g.). This minimizes data sent to distant cloud data centers, ideal for processing videos or sensor data from remote locations in public safety applications.In essence, edge computing brings the power of computation closer to the action, enabling real-time decision-making, mobility support, and efficient bandwidth utilization.Edge computing architecture primarily focuses on distributed computing resources across the network and includes three main components: edge devices, edge nodes, and the central cloud or data center.
ML
Semi-Supervised Learning
Artificial intelligence is just starting to resemble what science fiction writers described in their works. But this is already enough to open fundamentally new possibilities for us, which are now becoming an integral part of our life. And all these numerous possibilities are backed by various machine learning algorithms, which process data in different ways, depending on which models they were trained. Today we will dive into semi-supervised learning, how it works, what problems it solves, and what opportunities it provides. But let's remember the difference between supervised and unsupervised learning first.All this is explained in detail in our article, enjoy a beneficial reading.Supervised learning is an approach in machine learning where the model is trained based on labeled data, where each input example corresponds to a known correct label. The model uses this to extract patterns and create connections between input data and their corresponding labels in order to learn how to predict them for new, unlabeled input data.
ML
Recommender System Using Machine Learning
In today's world, recommender systems have become essential. They have made our lives easier and more comfortable, from suggesting books, movies, and songs to recommending what to buy next. According to reports, Netflix states that algorithmically-generated recommendations influence 80% of its viewership.But how do these systems work? What are the different types of recommender systems available? This article aims to introduce the various recommender systems and their strengths and weaknesses. What's important is that we discuss our personal experience of working with them.Recommender systems, also known as recommendation systems, are machine learning algorithms that use data to recommend items or content to users based on their preferences, past behavior, or their combination. These systems can recommend various items, such as movies, books, music, products, etc. Different approaches to building recommender systems include collaborative filtering, content-based filtering, demographic-based filtering, utility-based filtering, knowledge-based filtering, and hybrid approaches. The ultimate goal of recommender systems is to help users find items they will likely enjoy and increase user engagement with the application or platform.Making recommendations involves two main stages:Large enterprises such as Google, Instagram, Spotify, Amazon, Reddit, and Netflix frequently employ them to boost engagement with their platform and users. By collecting user data, Amazon utilizes recommendations to propose products to diverse users. For instance, Spotify suggests songs similar to the ones you've often played or liked to keep you using their service to stream music.
ML