FALCONY AI

Empower Your Business with AI-driven Solutions

About Us

We are a company specializing in a wide range of advanced AI, Machine Learning and Data Science technologies to support your business needs. Our expertise includes development around Large Language Models, Deep Learning, Graph Neural Networks, Classical Machine Learning, Recommendation Systems, Natural Language Processing, Data Engineering and ML Ops.
Additionally, our Algorithm Engineering focuses on designing, fine-tuning, and implementing solutions that perform optimally in real-world scenarios, both in terms of speed and the costs.
Partner with us to leverage the power of AI and unlock new possibilities for your business.

Zašto mi ?

Profesional Team

Fast Response

Quality

Services

LLMs for Businesses

Use of advanced language models improve customer interactions and automate various tasks. These models help you understand your data better. By utilizing these models, businesses can improve efficiency, personalize customer experiences, and make data-driven decisions with greater accuracy.

NLP

We combine our deep knowledge of Natural Language Processing (NLP) with valuable insights from the medical and financial sectors. This allows us to create custom language solutions that meet the specific needs of each industry. Our tailored approach ensures precise and effective results for your business.

Deep Learning

When traditional machine learning falls short, our expertise in deep learning frameworks like PyTorch and PyG steps in to bridge the gap. Deep learning uses neural networks with many layers to analyze and learn from large amounts of data. This approach excels at tasks such as image and speech recognition, natural language processing, and complex decision-making.

Recommender Systems

Boost your user engagement and conversion rates with tailored recommendation engines. Our recommendation system uses advanced AI algorithms to deliver personalized product suggestions, enhancing user engagement and boosting sales. By analyzing customer behavior and preferences in real-time, our solution ensures that each user receives relevant and timely recommendations.

Data Analytics

Unlock the potential of your data with our Data Analytics services. We turn raw data into actionable insights, giving you a deeper understanding of your business. This empowers you to make informed and confident decisions based on real data.

Data Engineering & ML Ops

Simplify and improve your data processes with our Data Engineering solutions. We create strong and scalable data pipelines for easy data collection, storage, and preprocessing. Our services are customized to meet your specific needs.

Algorithm Engineering

Let's talk about algorithms. We specialize in Algorithm Engineering, going beyond traditional algorithm design. Our expertise includes fine-tuning, implementing, and testing solutions. This ensures optimal performance in real-world scenarios, helping your business run efficiently and effectively.

Case Studies

LLMs and RAGs

While it is standard procedure to use RAGs based on text-segment embeddings, the use of Knowledge Graphs can significantly improve the accuracy of the chatbot. In this project, we leveraged our competencies in graph theory, algorithm engineering, vectorization, and LLMs.

Pricing Engine

Predicting a price at which a transportation carrier would accept transporting a load is not a trivial task. Besides distance, other factors play a role, such as weather conditions, seasonalities, and the nature of the load. For our client, we were able to design and implement a powerful pricing engine that reduced the error of the previous baseline by 60%.

Investments

We were asked to develop a system that would predict which hedge fund will invest in a particular stock and how many shares will be bought. Our solution was a Graph Neural Network model that was able to predict the links as well as the link weights (the number of shares). As a baseline, we used a graph-embedding approach based on Node2Vec.

Debt Collection

When collecting debt from bank clients, it is important to know the behavioral scores: how likely it is that this particular client will pay. Also, in the process of debt collection, one needs to know which is the best communication channel for a particular customer: whether to send an email, an SMS message, or to call the debtor. We were able to provide an effective and efficient solution, to our client's great satisfaction.

Anomaly Detection

When surveilling a network, one wants to know if a particular cell reading by a particular sensor was anomalous. Also, one wants to know if the signal quality and strength readings are anomalous, such that one can take appropriate measures. We tackled this problem via Deep Learning but also used Graph Neural Networks.

Recommender Systems

Recommending friends on a social network is a challenging task. A client asked us to develop a system that does dynamic user regrouping, assigning users to groups of like-minded people, which encourages in-person meetings. Not a typical Recommender System project, but one for which we provided a great solution.

Testimonials

"The Falcony AI team showcased strong technical expertise and a deep understanding of machine learning models and data science methodologies. Their work was integral in refining our technology, optimizing algorithms, and improving the system’s performance."
Elad Tzoreff
Head of Algorithms at Sensorz
"I highly recommend Falcony AI for their outstanding work in data science and machine learning. They enhanced our scoring model, improving its overall accuracy in production. Additionally, they developed a compliant way for our platform to respond to consumer inquiries automatically, without the risk of hallucinations.

The team demonstrated deep technical expertise. Their work had a tangible positive impact on our operations, improving our product offering."
Michael Beckes
CTO at Receeve GmbH
"Working with Falcony AI has been an exceptionally smooth and plesant experience. Their development of a graph-based predictive algorithm for StockFink not only enriched our predictive arsenal, but also showcased the seamless integration spirit of our collaboration."
Dr. Juan Luis Fernández - Martínez
Co-founder of StockFink
"Falcony AI has the great in-house expertise to let you navigate easily through your recommender dreams.
Mirza helped us to set up a detailed roadmap for our upcoming development."
Adrian De Bolle
Founder of UNBLND
"Navigating the AI landscape is challenging for any company today; that's why having the ability to do it with experts like Mirza and his team is so valuable! From turning complex concepts into understandable logic, over brainstorming AI projects with you enthusiastically with an eye for detail, to having the in-house knowledge and expertise to develop all your crazy AI ideas, falcony.ai has you covered! I cannot recommend enough the quality and willingness of this team to support you in successfully embracing AI and the future!"
Murielle Marie Ungricht
Founder of Nuumani.com
"Working with Falcony AI has been exceptionally straightforward and enjoyable. The team's work on feature engineering and improving our AI recommendation models significantly boosted our predictive capabilities."
Alfonso Gonzalez
Co-Founder of whizdom.ai

Our Team

Dr. Mirza Klimenta

Founder and CEO

Obtained his PhD in Computer Science from the University of Konstanz, Germany, at age 25. While in academia, Mirza worked in the fields of dimension reduction and graph embedding, and his work has been awarded by the scientific community. As a Data Scientist, Mirza focuses on Recommender Systems, Pricing Engines and Algorithm Engineering.

Jasminko Sofo, MSc

CTO

MSc with distinction in Computer Science from the University of Buckingham. 14 years of hands-on technical experience with the last 5 years in data related roles. Jasminko worked in the fields of insurance, IoT, telecom, proptech and fintech. An excellent strategist.

Majda Klimenta, MBA

IT Project Manager

A proven record of successfully leading and organizing IT projects. Her expertise includes meticulous planning, team leadership, and effective client communication, ensuring project success from start to finish.

Dr. Jusuf Ramić

Senior Data Scientist

Obtained his PhD in Mathematics from the University of Zurich, Switzerland. A strong background in Data Engineering and Data Science, with expertise ranging from NLP to leveraging proprietary knowledge graphs for insights. Gold medal at the Mathematical Olympiad of Bosnia and Herzegovina. Bronze medal at the International Mathematical Olympiad.

Companies we worked for

Lorem ipsum dolor sit amet, consectetur adipiscing elit. Ut elit tellus, luctus nec ullamcorper mattis, pulvinar dapibus leo.
John Doe
CEO
Lorem ipsum dolor sit amet, consectetur adipiscing elit. Ut elit tellus, luctus nec ullamcorper mattis, pulvinar dapibus leo.
John Doe
CEO
Lorem ipsum dolor sit amet, consectetur adipiscing elit. Ut elit tellus, luctus nec ullamcorper mattis, pulvinar dapibus leo.
Marcus White
CEO

Sectors we worked in

Conferences

01/07/2024

18.11.2024 - Building a Recommender Model for the Online Betting Shop Gamingtec

Sanida Redžepović

No Comments

Building a Recommender Model for the Online Betting Shop Gamingtec Recommender Systems integrated within an Online Betting Shop presents a particularly intriguing area of study: there is almost no research in this field, and devising an appropriate cost function to optimize poses a non-trivial challenge. In this talk, I will explore recommender models that cater to […]

26/06/2024

25.06.2024 - Graph-based Recommender Systems - when to use them

Sanida Redžepović

No Comments

Recommender System models continue to grow in complexity, yet some relatively straightforward approaches, which might provide better solutions for most businesses, are often overlooked. In this talk, I will delve deeper into recommender models based on graph theory and graph algorithms, ranging from PageRank walks to Graph Neural Networks. This exploration will include an evaluation […]

Our Blogposts

21/03/2024

RecSys for Beginners

Sanida Redžepović

No Comments

Why do we build Recommender Systems? Recommender Systems are central to nearly every web platform that offers things - movies, clothes, any kind of commodity - to users. Recommenders analyze patterns of user behavior to suggest items they might like but would not necessarily discover on their own, items similar to what they or users […]

20/02/2024

Using Graph Neural Networks for Predicting Investments

Sanida Redžepović

No Comments

How to leverage Graph Neural Networks’ solution for link prediction. Suppose you had information about the investments made by certain investors in the past few months. To make more informed investment decisions, you want to know which companies these investors are likely to invest in during the following months. You start analyzing the data but […]

We @ Podcasts

06/05/2024

From Algorithm to Application: A Look into the World of Data Science with Falcony AI

Sanida Redžepović

No Comments

Shownotes +++Unsere 25. Podcast-Episode: Vom Algorithmus zur Anwendung: Ein Blick in die Welt von Data Science mit Falcony AI+++ +++ English Version below +++ In dieser Folge begrüßt René seinen Gast Mirza Klimenta den Gründer von Falcony AI, der uns tiefe Einblicke in spannende DataScience-Projekte geben wird. Wir tauchen gemeinsam ein in faszinierende Themen wie […]

20/02/2024

Podcast Recomendations in the ARD Audiothek with Mirza Klimenta

Sanida Redžepović

No Comments

In episode 15 of Recsperts, we delve into podcast recommendations with senior data scientist, Mirza Klimenta. Mirza discusses his work on the ARD Audiothek, a public broadcaster of audio-on-demand content, where he is part of pub. Public Value Technologies, a subsidiary of the two regional public broadcasters BR and SWR. We explore the use and […]