Projects
Computer vision
A tool for recognizing data from video cameras to monitor construction.
Construction Process Monitoring
A video recognition tool for monitoring interior finish readiness.
Monitoring of interior finishing construction
Creating a program that recognizes certain types of documents (passport, driver's license, vehicle registration certificate, etc.)
Automatic Document Recognition
A service that effectively detects and predicts fraud at the contracting stage.
Identifying and preventing fraud in insurance
Improving the usability of the service and increasing the reach of users through the development of a telegram bot.
Telegram bot
Detection, segmentation and classification using training to identify bicycle parts.
Bicycles parts detection, segmentation and classification
A model for searching for structural damage on a small volume of images
Detector, segmentator and set of rules for determining damage
used open data with segmentation annotation from drone, build a segmentation model to find different classes: people, bicycles (22 classes)
Segmentation of objects for drone photos
Bounding boxes annotation, segmentation annotation, and classification annotation
Image annotation for CV project
- deep learning classifier for videos
- face search service with elastic search
- tagging for images implementation of a web service a site
Content video and photo AI system
A browser plugin that determines the quality of content on the site.
Content Quality Definition
A system for collecting a large number of video materials for subsequent monitoring of the construction process.
Collection of video materials to monitor the construction process
The software solution for removal of logos and watermarks from images.
Removal of logos from images
Scoring and analytics
Сustomize the formation of the optimal cost of the mortgage insurance product
Dynamic pricing for insurance business
- graph in python networks
- fraudulent customers were found in the graph
- results transferred to neo4j
- neo4j was developed to production mode
- GraphSage was part of the project
Graph analytics for financial company
Algorithms for constructing forecast values of frequency and the average loss in insurance.
Using Geodata in Predictive Insurance Models
Models to predict customer loss, loss size, and the likelihood of buying and renewing a policy.
Insurance company scoring
This technology can be used by companies when deciding on the purchase of building materials for the future.
Price forecast of building materials
Automatic collection of prices for building materials from well-known building stores
Parsing prices of building materials
- we have solved several optimization tasks for the retail industry (find a minimum cart with constraints for suppliers and buyers)
- it reduces the cost of buying from small companies
Optimized retail industry
- we have built SQL databases and OLAP cubes
- developed analytics based on OLAP technology (insurance portfolio audit, price change suggestions, analytic reports)
SQL and OLAP cubes development
- build model's deaths prediction (XGBoost, LGBM, optuna, XGBSE, pycox)
- helps to make life longer
Patients death prediction for hospital
- build models for metal prices (PyTorch forecasting, prophet, XGBoost, LSTM, greykite)
- make buy and sell recommendations for management
Time series prediction for commodities
The model collects and combines the necessary information from images and databases and acts as an assistant in protecting intellectual property.
Intellectual Property Protection
Multivariate portfolio analysis, identification of positive and negative aspects
Insurance portfolio audit
System to collect a video footage to record safety violations at construction sites
Collection of video to record violations at construction sites
- extracting data from different sources (articles, prices, content)
- automation of scraping and parsing process (airflow, cron)
- ETL process development and data storage
Web scraping
Creative industries
Creating a digital solution for a realistic cartoon that will reflect the signature styles of Silver Age artists.
Creating animations using artificial intelligence
Controlled generation of large volumes of images for training machine vision systems was required.
Synthetic image generation for machine vision system training
Development of an application that allows users to generate images online based on a given text query.
Generating images by text request
NLP projects
- Build the algorithm to generate articles from different domains
- Create web service with API
Article generator for SEO optimization
- we have built a URL articles parser with an automation process (airflow, cron)
- did different tasks for articles text (topic modelling, named entity recording, sentiment analysis, spam classifier, mistakes, summarization)
- build meta-model for text classification
NLP web articles classification problems
- build a big inner email communication graph
- did different tasks for email texts (topic modelling, named entity recognition, sentiment analysis)
- extract non-formal groups in a company based on topics and entities
Different NLP tasks for Oil and Gas company
Our servers available for free on projects
IBM x 3960
1TB RAM
120 CPU cores
3 GPU 1080ti
120 CPU cores
3 GPU 1080ti
Our servers available for free on projects
IBM x 3960
1TB RAM
120 CPU cores
3 GPU 1080ti
120 CPU cores
3 GPU 1080ti
120 CPU cores
1TB RAM
3 GPU 1080ti
Social projects
The company's team of interns participated in the largest hackathon of the year and won second place in their chosen category.
Leaders of Digital Transformation 2023. Participation in the Moscow Mayor's Contest
More than 50 research videos (data science, front-end, back-end, product management)
Data science You Tube channel
- developed and launched the educational course “Machine Learning in Business”
- trained over 150 students
Training course for students and interns of the company
