DataJaala — the groovy AI Lab

In mid 19th century Jaala ship was developed in coastal cities around the Gulf of Finland. The Jaala ship had originally two or three mast with gaffsails and the very latest Jaala ships had also motors installed in mid 20th century. The Jaala ships were sturdy and reliable cargo vessels providing growing cities with firewood, sand for construction and other useful materials and goods. Night or day, storm or good weather, the Jaala ships delivered valuable cargo for customers.

In spirit of Jaala ships, the DataJaala's mission is to provide high quality machine learning and software engineering services to customers with impeccable execution and a good dose of humour and creativity. DataJaala wishes to be Aidosti hyvä juttu.

What makes DataJaala different?

Unlike many other larger consulting companies, DataJaala is restricting itself to only certain areas and problems. DataJaala personnel have in the past worked in web projects, control systems, automation systems, customer analytics, churn prediction, time series analysis, time series classification, anomaly detection, datapipes, cloud data platforms, on premise and so forth.....and the key lesson has been that if you want to be really good at something, then you have to pick a speciality. For DataJaala the topics of choice are time series technologies, time series prediction, anomaly detection and MLOps supported by extensive cloud experience.

However, DataJaala's long and extensive experience allows working in all roles (data engineering, algorithms...) in various machine learning related projects within our speciality. In addition, DataJaala keeps track of latest developments in AI literature and has experience in implementing large scale cutting edge machine learning systems making DataJaala very versatile and effective partner.


Services

DataJaala provides machine learning consulting services on an hourly basis and has extensive competence and experience with:

Machine learning on time series data
Prediction, forecasting, analysis, anomaly detection, classification and so forth.
Time series technology
Time series databases, processing, streams, ETLs, efficient breakpoint transformations, filtering, gap filling and so forth.
MLOps
AWS Sagemaker, Databricks, Azure Machine Learning and also model serving infrastructures built from ground up.
Cloud
Long and varied experience with cloud platforms with Infrastructure as Code and automated provisioning.

Predictive maintence kick-start

Predictive maintenance kick-start will help you to get started in predictive maintenance journey by identiying early use cases and by creating predictive models based on your data.

R&D can be done DataJaala secure cloud analytics environment or in customer systems depending on the customer preference.

Price: Hourly rates apply

MLOps architecture

MLOps is indispensable part of any professional machine learning practice. To get started on the MLOps journey one needs to have clear overall MLOps architecture and vision.

In productized MLOps architecture project DataJaala creates MLOps architecture in collaboration with client domain experts and data scientist.

Price: 19 500€ (VAT 0%)

Coding Data Platform Achitect

DataJaala Coding Data Platform Architect helps you to create practical data platform architecture and to lead platform implementation work on both team leading and in actual implementation aspects.

Price: Hourly rates apply

Coding Lead ML Engineer

DataJaala Coding Lead ML Engineer consultant helps you to create advanced machine learning models and to deploy them to production to create practical business value.

Price: Hourly rates apply

Consultants

Ilpo Suominen did his first datascience project in 2005. After stint (and Doctor of Science, Technology) in the fascinating world of computational physics and electronic structure of high temperature superconductors, Ilpo has been working in technical roles in startups, consulting companies and stock listed corporations.

Ilpo's CV is available at Linkedin profile and his a bit more freewheeling ideas and approach to datascience and software engineering is available at his blog


DataJaala Lab

Datajaala Lab keeps up the latest research in anomaly detection, time series and generative models. In addition to this DataJaala Lab tries out selected models by implementing them to investigate practical usability of the model in industry. DataJaala Lab also investigates key technologies needed in machine learning, big data and cloud services.


Stateful spark streaming with deltalake

In breakpoint time series only changed values are indicated and therefore the data must be to transformed to equally sampled data before using in most time series algorithms.

See project github repo here!

LSTM anomaly detection

LSTM sequence to sequence model anomaly detection implementation based on LSTM-based Encoder-Decoder for Multi-sensor Anomaly Detection.

See project github repo here!