Sparkbit

CHALLANGE:

Happy Team Check co-founders knew exactly what they wanted – a Saas platform that transposes employee satisfaction surveys used by companies like Facebook and Google – and converts it into a great tool for a selected industry sector. Working closely together, we came up with solutions tailor-made for our client.

A web portal for conducting flexible employee satisfaction surveys, targeted at small and medium-size enterprises. For Happy Team Check we developed a system that defines and conducts customized employee satisfaction surveys. 

SOLUTION:

The system’s backend is written in Scala and Slick while frontend in AngularJS. It includes a database of predefined questions that were created and shared by HR experts.
Yet each client can define additional questions.

Many types of questions are supported, such as free text, single and multiple choice or Likert scale questions. 

IMPACT:

We delivered a tool that helps businesses to maximize the potential of their employees.

CHALLENGE: 

Sparkbit works with Toolbox for HR to create a user-friendly platform to keep and track data of candidates and campaigns within the system. The goal of the project was to provide recruiters with an easy time-saving tool. This application would eventually remove the need for spreadsheets and cut down on the flow of documents in the office.

Fermi supports customizable workflows and tracking of a recruitment process. 

SOLUTION:

It facilitates overview of the candidate’s number during each stage and control recruiters’ appointments. Fermi includes an advanced search engine.

Recruiters are able to filter data by candidate, campaigns and recruitment process stage. 

CHALLANGE:

Our challenge was to create a user-friendly application which will help to effectively conduct interviews, track data of candidates and campaigns, analyze collected data and compare candidates.

The goal of the project was to provide time saving tool gathering all of the informations in one place and eventually replace the paper-based recruitment process.  

Toucan is a comprehensive and user-friendly tool for taking control over recruitment process. 

SOLUTION: 

Toucan supports the recruitment process end-to-end: it can automatically import candidate profiles (including all attached documents), allows to preview the profiles, schedule and conduct interviews. It also provides user friendly statistics and KPIs. The interview mode is optimized for a tablet and supports creating hand-written notes (both by recruiter and candidate) that are securely stored in the system.

With Toucan, paper flow is entirely eliminated, sensitive data is stored securely and remains accessible for future recruitments. 

IMPACT:

We delivered an intuitive and time-saving tool that helps recruit faster and easier.

CHALLENGE:

Timepot MVP released in January 2017, covered basic functionalities of the application – time tracking and reporting.

Having MVP already published we focused on building an advanced financial module that compares various projects’ profits. Its algorithm calculates monthly income and expenses of projects run in different paying models, several currencies with mixed employee’s wages. 

SOLUTION:

The system’s backend in written in Java 8, Spring while frontend in Angular 4 and TypeScript. Timepot offers different access levels attached to specific functionalities of the system.

Its reporting module includes an advanced search engine.

Users are able to filter data by projects, tasks, keywords and users. The data is displayed on clickable charts and table and can be exported to CSV and PDF files. 

IMPACT:

We delivered a beautifully designed and intuitive tool with powerful functionalities that gives the clients insights into their businesses.

CHALLANGE: 

Technical or scientific documents usually have a lot of figures, tables or diagrams besides text. In one of our systems it was crucial to extract these kind of objects from a PDF document as a separate image and take note of its location in the original document

We have used deep learning techniques to identify objects of interest in the documents.

SOLUTION:

We have implemented the YOLO detection algorithm and evaluated multiple architectures of the underlying convolutional neural network.

We have pretrained the model on a large set of labeled images and then applied transfer learning to our specific task.

To increase the data set for our specific task, we have applied data augmentation. 

IMPACT: 

Using Deep Learning mechanism we created system which can detect and extract objects from fixed-layout documents.

.

CHALLANGE: 

Italist is an e-commerce platform offering 1,000 international brands and 200,000 luxury items for customers from all around the world. It’s a system that has millions of visitors annually. It was originally built as a PHP monolith powered by MySQL database.

Over the years the number of users and features had grown to the point where the system became no longer maintainable nor extensible.

The goal of our client was to broke one monolith system into smaller, more manageable subsystems.

SOLUTION:

Our client decided that the system needs to be broken into smaller, more manageable subsystems. This also involved containerizing each new subsystem so they could be easily deployed and scaled.

We joined the core Italist team in an effort to make this transition go smoothly (we did it by minimizing potential regressions) and quickly (we were time constrained by the holiday season).  

IMACT:

We've made the transition from one monolith system into smaller more manageable parts go smoothly and quickly.

CHALLANGE:

The classic pension systems most people are currently a part of are inefficient and are not align with the current digital generation. There are places in the world where such systems cannot be trusted as the governments can seize the funds at any moment. The goal of our client was to tackle those issues by creating a platform for ordinary people and pension funds to use with trust and efficiency at the heart of the design. The goal of our client was to create a secure platform for ordinary people and pension funds which will be based on blockchain

SOLUTION: 

A natural fit to this challenge is blockchain. By using smart contracts implemented in Solidity on the Ropsten Testnet we were able to create an MVU that demonstrated the basic concepts of the future full implementation of the plaftorm. The protocol itself used an utility token for fees.

On the Frontend side we implemented a React dapp that allowed the user to choose a pension fund and invest ETH in them or buy utility tokens. The development process included a full suite of tests for the smart contracts and CI that would deploy the contracts on a test Ethereum-compatible blockchain.

IMPACT:

We have helped to create a secure pension infrastructure based on blockchain.

CHALLENGE: 

Our client was creating a security analysis SaaS platform for smart contract projects. Users of this system would subscribe to paid plans that would allow them to use automated audit tools for their code. The envisioned payment method for this service was a utility token deployed on Ethereum. This token would be exchangeable for Ether by using a market contract utilizing a bonding curve price function. This would eliminate the need for third-party crypto exchanges and reduce speculation on the price of the token. Sparkbit was contracted to create a web application handling the subscription process and to implement the smart contract code that would handle the payment processes and the market contract.

SOLUTION: 

Both the dapp and the contract code integrated with the clients API to provide a seamless experience for the user. In the near future the bonding curve pricing model will be validated against possible abuses using Mutli agent reinforcement learning techniques. The smart contract will be audited by an external auditor and deployed on mainnet Ethereum.

IMPACT: 

A working subscription frontend and smart contract backend was delivered to the client.

CHALLENGE: 

Our client needed a platform for management of a next generation investment fund. The fund is denominated in tokens on the Ethereum blockchain, invests into other cryptocurrencies and was distributed through an ICO. The system should allow for configuration of the investment strategy, withdrawal of the funds to user’s blockchain wallet and displaying the current fund structure and user’s share value. The system had to fulfil security-related regulations, thus the identity of the user had to be verified and the ownership of a blockchain wallet had to be confirmed.

SOLUTION: 

Sparkbit was hired to design the architecture and implement the entire system. We have delivered a web-based platform that is integrated with the blockchain, external APIs with financial data and internal systems of the client. We have also helped the client to identify and resolve some security issues in their initial requirements.

IMPACT: 

After a successful delivery of the first stage of the project at the end of 2017, we are now working on next phases.

CHALLENGE: 

Our client wanted to include in his offering a telematics platform that would be offered to fleets. The goal was to evaluate driving behavior of users based on data collected from vehicle’s OBD2 interfaces as well as a GPS and accelerometer.Our challenge was to create a horizontally scalable platform that could be also extended to handle other sources of data in the future.

SOLUTION:

We have implemented a system based on big data stack: Kafka, Scala, Spark Streaming and Cassandra. To simplify portability between different cloud providers, the system is dockerized and Kubernetized. The platform also utilizes Ethereum blockchain technologies to transparently and reliably store some vital data of the system.

IMPACT:

We have created a telematics platform with blockchain component that would be offered to fleets.

CHALLENGE: 

Our client performs online surveys for psychological risk and work-related stress assessment for German companies. He needed a tool which will be used for automated analysis of the survey’s results and knowledge extraction from the available data.

Working closely with the client we’ve turned his needs and vision into the functioning product.

SOLUTION:

DearEmployee is a platform for analyzing and depicting data gathered from psychological surveys.  Its most important feature is providing in-depth insight into the employees’ health and relationships as well as internal metrics such as absenteeism and employee turnover.

Results can be filtered by business units, areas of activity and other target groups. There’s a granular access control - different people have access to different parts of the results depending on their role in the organization.  

IMPACT:

Working closely with the client we’ve turned his needs and vision into the functioning product.