Data Science @dataeaze
What we do ?
Helping companies build data-driven products,
by leveraging cutting edge AI technologies
Dataeaze systems is focused on making it easy for organisations to work with data. We help organisations build "data driven" software products, by incorporating domain knowledge and leveraging machine learning. With artificial intelligence and deep learning techniques, we build "software 2.0" data pipelines for all kinds of data -- streaming or big data, structured or unstructured.
We ease process of building Data Driven Products
dataeaze - core expertise (techniques)
Data Management
data analytics
data visualization
data mining
data pipelines
Core Computer Science
data structures
graph and networks
databases
operating systems
Machine Learning
kernel methods
forests, ensembles, boosting
clustering
collaborative filtering
dimensionality reduction
Artificial Intelligence
natural language processing
information retrieval
speech recognition
machine translation
Systems / Computing
distributed computing
GPU
cloud computing
edge computing
embedded devices
Software Engineering
software development
technical evaluation
production deployment
ML software lifecycle management
dataeaze - core expertise (tools)
Data Management
Core Computer Science
Machine Learning
Artificial Intelligence
Systems / Computing
Software Engineering
Data Science Workflow
Ongoing projects
Deep Learning powered Intrusion Detection
We have designed and developed the AI/ML software stack for a intrusion detection solution for a defence application. The hallmark of the solution is the application of computer vision and deep learning techniques to spatio-temporal signals to a completely different domain -- fiber optics.. In doing so, we have built possibly one-of-its-kind "third generation" artificial intelligence system for this problem -- where the need for feature engineering is eliminated.
Area - fiber optics, deep learning, computer vision, defence
MLOps
We developed a DevOps toolkit for machine learning. For a machine learning engineer, it offers automl with hyperparameter tuning as well as experiment tracking. Code, model and data versions for all experiments can be tracked. It provides minimum code change workflow along with visual "big picture" view, while not compromising on scalability and applicability to wide set of machine learning algorithms.
Area - MLOps, productivity, automation
Machine learning for telecom
We are applying machine learning techniques to build a software product that aims to reduce the time and efforts spent by NOC engineers in diagnosing faults in a telecom network. The software helps telecom operators to improve service assurance and helps the telcom service providers to reduce their operational costs
Area - machine learning, telecom
Upcoming projects
Topic Modeling
Automatically discover topics from a corpus
Area - Natural language processing
Weapon Detection
Detect people carrying weapons
Area - Computer vision for security
Active Learning
Human in the loop system which reduces labeling efforts
Area - Productivity in Machine learning
TryOn
System to try on clothes virtually
Area - Deep learning for computer graphics
Search
Search engine for images, text, audio etc
Area - Information Retrieval
IoT
Data engineering stacks for IoT particularly for automobile manufacturers
Area - IoT, federated learning
Music generator
App for generating music ringtones
Area - Generative Adversarial Networks for audio generation
MLOps
Managing life cycle of ML products and ML experiment tracking. Track model, code and data versions in a consolidated view
Area - MLOps
Car Damage Detection
Car damage detection for insurance companies
Area - Computer vision applications for business
Security
Multi camera system involving person reidentification, action recognition powered with jetson nano.
Area - Privacy preserving AI for surveillance
Handwriting Recognition
Handwriting recognition from scanned documents
Area - Document Imaging
Traffic
Counting, event alerts, statistics, etc on traffic videos
Area - Computer vision for smart cities
Intrusion Detection from Thermal Images
Person detection from thermal images
Area - Intelligent Sensors
Testimonials
Emergent technologies in fiber optics have brought in a new set of engineering problems that demand for artificial intelligence/machine learning approaches. Dataeaze helped us to tap the power of computer vision and deep learning techniques for the modeling of one such engineering problem. Following a rigorous data science and engineering discipline, they built a software stack for us. It was a complete ''software 2.0'' lifecycle execution, from ideas in the research lab to the code deployable in production. I wish them success and hope to work with them again in the future.
I had the pleasure of working with Dataeaze over the course of a year - their passion for AI, deep technical expertise and focus on execution helped us to build a complex deep learning solution from ideation to deployment in 8 months. They take full ownership of the pipeline and understand the complexities of production. Ajit is a mature leader of his team and this shows in their ability to retain world-class talent.
The product was delivered within a reasonable time. The Dataeaze team proved to be flexible when faced with challenges and project unknowns, coming up with solutions and new ideas. They reported on a weekly basis and used a transparent project management style.
Clients and Partners
Work with us
Join the team of highly professional developers
Connect With Us
Dataeaze Systems Private Limited
Plot no 41, Lane no 3, Shivaji Housing Society, Senapati Bapat Rd, Shivajinagar, Pune, Maharashtra 411016.