Machine Learning
Our engineers have helped companies across industries leverage Artificial Intelligence and Machine Learning technologies to scale their business, optimize their processes, achieve their engineering initiatives, and deliver high performance to their customers.

Artificial Intelligence and Machine Learning projects
reduction in work needed as reported by one of our clients, resulting in time & cost savings
reduction in manual work due to increased efficiency, as reported by another client
What we do
Computer Vision
Understand and interpret the visual world, generating AI visual art
Natural Language Processing
Develop intellectual agents that can handle natural languages, closely simulating human communication
Time-Series
Build AI systems that can process time series data. These systems are used in finance, healthcare, and other industries
Signal Processing
Create AI systems that can process signals such as audio, video, brain waves, and more
How AI & Machine Learning can benefit businesses

Save time and reduce costs by automating processes

Improve business value by predicting customer behavior, allowing to deliver preemptive care
Find insights by processing and interpreting immense amounts of data

Protect against risks and mistakes by reducing the chance of 'human error'

Our Partners







AI & Machine Learning step by step
Understanding Business Goals
- Understanding business processes
- Identifying problems and growth areas
- Acknowledging and preparing for all potential risks
- Defining success criteria
Data Exploration and Features Engineering
- Identifying sources of data
- Collecting data
- Integrating data obtained from different sources
- Preparing and cleaning data
- Performing data augmentation
- Crafting features using domain knowledge
Building ML Model
- Model selection and validation
- Model training
- Model testing
Delivery
- Deploying the model
- Monitoring the model
- Evaluating new data
- Adding extra functionality
Support
- Monitoring
- Maintenance
- Continuous improvement
Our AI & ML expertise map per domain

Akvelon AI & Machine Learning projects
Classification

Advertising Campaign Evaluation
Delivered classification of user actions and attribution matching to predict the value and ROI from advertising campaigns, and then compare it with campaign costs.

Cat Translator App
Developed audio classification for a cat vocalizations translator using Machine Learning to operate on mobile devices. Used the YAMNet acoustic detection model (converted to a TFLite model) with transfer learning.

Legal Documentation Automation
Performed entity extraction, text matching, tabular data classification, text similarity measurement, custom ML models development, categorization of legal documents, and legal conflicts searching.
Segmentation

Interactive Whiteboard
Created an interactive voice-controlled assistant for ad-hoc whiteboarding sessions to save all records made. Used a webcam, an edge detection algorithm, and human body segmentation with U-Net architecture.

Smart Parking
Developed a recognition service to automatically detect open parking spots at busy parking lots that do not utilize painted parking lines.

Warehousing Optimization
Tracking assets location using machine vision seems. Created own unlimited synthetic training datasets with Unity 3D and used U-Net segmentation models with different backbones (MobineNetV2, EfficientNetb0-4) for software.
Regression

Marketing Campaign Predictions
Performed sequence prediction of the number of installs and total value of networks. Delivered a prediction of the clients’ and users’ Long Time Value by groups based on aggregated matching and event model for advertising and marketing campaigns.

Customer Satisfaction Analysis
Developed a deep learning model for analyzing customers’ attitudes in real-time via onsite cameras. Built using transfer-learning and expanding on a pre-trained model that was designed to detect facial expressions during trade shows, live demos, and in-store.
Clustering

Targeted Online Advertising
Performed Bing search queries clusterization for our retail client, clustering their customers based on their product preferences for optimized and personalized online shopping experiences. Generated exciting and unique deals for users.

Linkedin Research Optimization
Trained a classification model and clusterization to split LinkedIn users by groups according to their job description in order to optimize research on individuals and companies. Built n-dimensional vectors based on company description, and grouped companies by similarity.
Anomaly Detection

Time Series Anomalies Detection
Developed a statistical model and neural networks to detect and predict variations (anomalies) in time series and textual data. Provided evaluation with static and dynamic threshold on a scarce training data set of log metrics in order to enable full monitoring for all clients of a web app.
Generation

AI-Powered Smart Filter App
Developed picture processing and generation models for an application and web platform that transforms photos and videos into digital art using neural networks. Developed smart filters that replicate artistic styles of various artists.