Big Data and Artificial Intelligence (AI) are transforming global businesses and are significantly contributing to economic growth via contributions to productivity. These disruptive technologies are addressing some of the biggest challenges in critical areas from healthcare to climate change. In an exclusive interaction with Analytics Insight, Anton Shaleynikov, CEO and Founder of Dashbouquet Development shares how AI is becoming the core component of an effective digital transformation strategy and how the technology is creating opportunities for businesses and societies.
Kindly brief us about Dashbouquet Development, its specialization and the services that your company offers.
Dashbouquet Development is a frontend development agency, located in Belarus. Our clients are ambitious startups with big ideas, and established enterprises who are looking for advanced solutions. We help startups to build their big things and we help to get corporate software products from the stage of concept to production.
We started in 2014 and have completed dozens of projects since then, with the majority of them going into production. We do a full development cycle: from prototype and designs to final testing and market launch, as well as production support. Our team is composed of 40 people and we have two departments. In addition to web development services, we have a product division that develops our own applications.
We are technology savvy and have a strong focus on JavaScript and Data Science. Such an approach grants us profound and deep knowledge of the technology that we own.
With what mission and objectives, the company was set up? In short, tell us about your journey since the company launch?
I founded DashBouquet Development in September 2014. Before that, I was running a few tech startups and already had experience in the purchase of several IT services. However, almost all of them didn’t meet my expectations in terms of timing, flexibility, and proper communication. So, the idea was to create a company that would match these criteria and would be a good business partner.
Back in the days, we started with a small team of 5 people, and even though we did not have extensive experience, we had a clear vision of how things should be. We found our first clients on elance.com (now UpWork) and quickly achieved a position in the top 30 out of ~58,000 companies represented there. Our priority has always been the building of a reliable partner network in order to have good references. Such an approach has already brought us a lot of good projects and it keeps working. 30% of our clients come back to us with the next project.
Please mention some of the awards, achievements, recognition and clients’ feedback that you feel are notable and valuable for the company.
DashBouquet was recently ranked as one of the top web developers in Belarus and a top app development company in Belarus for 2018 on Clutch – a leading business-to-business service research and review company. Since our foundation, we have served companies around the world and truly have a passion for what we do.
We tend to focus on a few specific technologies and master them. Such an approach allows us to excel our skills and knowledge and results in great feedback from satisfied clients. We forge the partnership based on trust the way it should be.
Tell us how the company is contributing to the AI/Big Data Analytics/Robotics/Self-Driving Vehicles industry of the nation and how the company is benefiting the clients.
The project, that is already in production, is aimed at implementing Machine Learning and Text Analytics technologies into the HR market. The client came to us with an idea to make the recruitment process more efficient and simpler.
The value proposition is to help recruiters focus on interviewing the right candidates instead of screening thousands of resumes in order to get the perfect match. The employer won’t need to collect resumes manually from multiple platforms. The only thing he/she will have to do is to choose a number of platforms, and the system will collect all resumes and transform them into a readable format.
The entire system is fully flexible and allows to configure publications of job openings and resume admissions based on certain requirements and financial possibilities. With such a program, any company will have an opportunity to customize its hiring workflow in accordance with the company’s needs.
Machine learning allows screening applicants by their requirements matching level, using the matching engine. At the same time, the text analytics technology automatically converts job description text into data and the applicant’s resume files (Word, PDF and etc.) into searchable data. The implemented algorithm learns the employer’s recruitment patterns and improves the matching results based on that.
In order to make the screening process easier, we implemented the algorithm that automatically determines the candidate’s compliance with the vacancy on a scale from 0 to 10. This algorithm allows automatic rejection of all candidates with a low level of compliance.
What we are currently working on is a group of applications that would interact with a user through a webcam. The camera detects people and certain points on the body. For example, it detects eyes, so that it will know where to put on the glasses. The second part of the algorithm is detecting gestures. For example, there is a hand movement. When you are waving, something is happening on the screen. And specifically, the program can detect spread arms like the “plane” pose. The system also identifies age and gender.
For the upcoming winter, we are working on a special Christmas feature – the drawn falling snow stays on the head and shoulders and then crumbles, without flowing through a person.
As an outcome of this interaction, the users get a promo code or a photo to share with their network.
Technological stack for such an app includes Javascript and React for code, Python, OpenCV, and Tensorflow for detection and recognition, while algorithms for physics and interaction with the objects are pure mathematics.
We are constantly expanding our technical expertise. For now, we have three independent projects on kaggle.com. The first project is about recognizing the poker hand without specifying the game rules. We used a fully connected neural network and feature engineering and got 100% accuracy. The second task was photo recognition of dogs’ breeds. We used three models: ResNet50, Inception, and Xception. As a result, we ended up with 85% accuracy, which was lower than we expected. But we see it as an opportunity to determine mistakes and increase accuracy for the next iteration. The current task is photo recognition of automobile brands. For that, we use the Siamese Neural Network. The project is live and we are waiting for the outputs.
Please brief us about the products and services you provide to your customers and how they get value out of it
We develop mobile and web applications; particularly, we offer expertise in computer vision and data science. We take full responsibility for the project and construct it “from scratch.” As for the after-production support, we always help our customers a.s.a.p., including urgent issues, overnights or holidays.
What we always try to do is focus on what matters, analyze everything, and keep the communication with the clients ongoing. Our main principle is: “demonstrate instead of telling a story.”
We prefer to check everything ourselves rather than relying on someone else’s opinion, and we constantly decrease losses in communication by getting closer to a problem.
The best advice I can give is: “Always listen to your clients and don’t try to tailor them to some standards. Every client has individual requirements, so your prior task, as a developer, is to understand what they want and together come up with the best solution.”
The whole web development arena is moving toward possible user needs and tries to be proactive instead of reactive. There is a lot to explore and a lot to implement. The creativity of developers also plays a major role in “likeability” of an app. We keep an eye on the trends and make sure to try them. We can confirm that we use innovative technologies, and we try to deliver the best product and experience to our clients.
What are the key trends driving the growth in Big Data analytics/AI/Machine Learning?
AI is one of the core components of an excellent digital transformation strategy. And, as more companies are vying to implement digital transformation strategies, it will grow more central in web development than ever before.
Gartner estimates that the total revenue of the AI industry will experience a 70% giant leap increment by the year-end, reaching over $1.2 trillion. It was also estimated that in 2019, more than 40% of organizations in the world will deploy AI technology to enhance productivity and customer satisfaction. These are all clear indications that AI will emerge as a leading trend in web development in 2019.
If we take a look at the industries, i.e. healthcare, it is a broad field. With AI and Big Data, a scientist will soon get the opportunity to prevent certain diseases, like cancer. This can be done by analyzing a patient’s history and all their records, so AI will be able to understand the mechanism of disease, enabling doctors to be proactive instead of reacting. Secondly, there are more and more intelligent prostheses popping up – this also offers a whole lot of opportunities for both doctors and patients. Such prostheses are able to distinguish images, light, pressure, etc. and are already a great breakthrough in the healthcare industry. And, of course, there are robot assistants. They can come in many forms – virtual nurses in the smartphone that are able to recognize warning signs and call for help or real robots that can monitor patients, alarm doctors and send notifications in case there is such a need.
Talking about the financial sector, it is well-known that Machine Learning and AI are great when it comes to Big Data analysis and performance of routine tasks by the machine. AI minimizes the chances of human error and is able to build predictions based on the historical data – these predictive models are used in predictive analytics when one wants to prevent fraud, for example. predictive analytics helps companies minimize future risks, estimate development, build strong and efficient strategy and much more.
Based on our previous and current projects I can say that our clients are willing to transform processes with AI bringing it to all industries.
