Upp Agency
Your AI & Data partner
We provide a complete range of AI and data services to power solutions for your toughest challenges.
We build impactful solutions.
We help companies design, build, and deploy AI and data solutions. We have a track record of 7 years in empowering businesses to drive innovation and create significant impact.
Shaped by cross-industry experience.
From startups to multi-billion dollar enterprises across 8 industries, our experience allows us to look at challenges from many angles, understand the big picture, and transfer effective strategies across sectors.
Powered by full-stack expertise.
We are dedicated to a holistic, full-stack approach to problem-solving. We excel at both strategy and execution, cover diverse AI domains, and our full-stack data scientists and engineers can tackle an exceptionally high range of challenges.
our services
Complete AI & Data Solutions For Your Business
Data Engineering 02
Operationalization 03
Previous projects
Recommendation system platform
AI solution for the hospitality sector designed to increase hotel and resort guest spending while providing guests with a personalized experience. The solution was powered by Recommendation systems and was integrated with several apps and touchpoints within the hotels to empower hotel staff suggestions and create on-property upsell opportunities smoothly and unobtrusively.
Demand prediction &
inventory optimization
Google-backed platform for fine-tuning the supply chain and inventory management across stores and warehouses, with goal to improve profitability and reduce waste. The platform was powered by optimization algorithms and demand prediction scaling at several 100,000 SKUs, providing substantial ROI for clients ranging from multimillion to multi-billion-dollar enterprises.
Yield prediction and crop placement optimization platform
Intelligent platform for farm enterprises that provides insights into current crop conditions, yield estimation, and crop placement suggestions based on previous results and business constraints. Models were powered by predictive ML for yield estimation and optimization for crop suggestions, and various data types were utilized, including satellite imagery, crop & soil attributes, laboratory results, and others.
Churn prediction &
customer lifecycle management
Intelligent system for estimating customer churn likelihood and customer value scores. Powered by predictive ML, it provided key information to marketing teams, enabling them to prioritize and refine churn prevention efforts, while also offering clarity on the reasons behind customer dissatisfaction, making this information accessible to a wider audience within the organization for informed decision-making.
Automated data extraction
and deduplication
A set of NLP-powered solutions for automated extraction of relevant data from images (Optical Character Recognition & Named Entity Recognition), and deduplication of entities (Entity Matching). Solutions were designed to remove manual data extraction from real estate flyers and to increase data accuracy within large databases for clients offering Data as a Service solutions.
Data architecture design & implementation
Designed and implemented a robust data architecture. Automated the collection and standardization of data on a large scale. Developed data pipelines for a wide range of data sources, including both bulk and streaming data, as well as structured and unstructured formats.
Occupancy prediction and dynamic pricing app
An intelligent tool that provides suggestions to hotel managers for adjusting prices based on room demand, aiming to increase hotel profit and reduce time spent underbooked. A combination of ML and traditional statistical methods was used to reach good levels of occupancy estimation.
Remote sensing object detection and analytics
Research & innovation EU projects for regional automated mapping and classification of agricultural land types, and estimation of biomass & carbon dioxide stored in forests. Collected and processed large volumes of satellite imagery and created Computer Vision models for imagery and Machine Learning models for estimations.