DataRoot Labs

Please describe your company and your position there.

Founder of a small smart-up software company developing an ML-based plugin for consumer email.

For what projects/services did your company hire DataRoot Labs, and what were your goals?

We had collected a lot of data, but frankly I was not sure whether anybody would be able to make a machine learning model effectively out of it. I interviewed a few different ML-specialized development firms and provided a sample of the data. My goal was to receive an accurate classification model based on the proprietary data I had collected.

How did you select this vendor and what were the deciding factors?

The factors were

1. specialization in ML, as opposed to other firms that claimed to be able to develop a solution in ML but did not specialize in it
2. a further sub-specialization in NLP
3. a random personal connection to the COO (we had an acquaintance in common)
4. strong interactive demo that was presented to me as part of the proposal.

Describe the scope of work in detail, including the project steps, key deliverables, and technologies used.

1. Development of automated data cleaning scripts
2. Experimentation of various NLP models to find the most accurate ones using Python.
3. Once a model was selected, tweaking of parameters to make the model as accurate as possible.
4. Development of a classification API using the final model. (Node / JS / TS)
5. Transfer of the API, Training scripts, and documentation to our Github / cloud services environment (GCP / Kubernetes / CircleCI).

How many people from the vendor's team worked with you, and what were their positions?

I think three. One was the 'lead' who developed the model. Another was the data engineer who developed the pipeline. A third helped out on devops to transfer the model over to our repo.

Can you share any measurable outcomes of the project or general feedback about the deliverables?

DRL achieved beyond what they had promised, delivering a model that was above 95% accurate according to the data I provided. Their delivery of the API allowed us to implement the model they created within days, which made for a very clean handoff. They provided analysis of the categories which had been provided and included documentation for training the model in the future with different data. All in all, they understood the task very quickly with very little explanation, delivered exactly what was asked efficiently and professionally.

Describe their project management style, including communication tools and timeliness.

We created a joint slack room to share updates and questions. They were quite responsive and were timely and always willing to get on a zoom call.

What did you find most impressive or unique about this company?

They understood the goal of my project quickly and we were on the same page from the beginning.

Are there any areas for improvement or something they could have done differently?

Not really.

Please describe your company and your position there.

I am one of the co-founders and CEO of SavStart. SavStart is an Artificial Intelligence powered platform that collects, analyses and contextualizes data that encompasses all things business. It’s a comprehensive tool allowing individuals to work on their amazing ideas with a competitive edge. 

For what projects/services did your company hire DataRoot Labs?

Our team comes from a background of business, technology and design. We came up with the idea to build this platform to help other business hopefuls. In the beginning, it was like trying to figure out how to build a time machine, it almost felt like mission impossible. Ivan and his team at DRL helped us turn our "time machine" into a reality.

They understood exactly what we were after and within a matter of weeks, they added all the parts needed to make our idea a reality.

What were your goals for this project?

We needed confirmation that our idea was something that could actually be built into a real working platform. The DRL team not only understood the technology needed to make the idea a reality, but they took the initiative to fully understand all aspects of our idea and the future goals we have.

How did you select DataRoot Labs?

Our research for the perfect team to help us develop our platform took us out of the UK. We searched for companies all across the world and after narrowing down to 5 which we went and met personally, we knew instantly that DRL and Ivan their CTO was the team we would want to work with.

They immediately had the best understanding and a portfolio of work to prove they knew what they're talking about, especially when it came to artificial intelligence.

Describe the project in detail.

After the initial face to face meeting, we had additional online meetings with Ivan. We also got introduced to more of the DataRoot Labs team. From the first discussion to getting a breakdown of what technology we would need, along with pricing and recommendations, the process took no more than 3 weeks.

From that point onwards the DRL team has kept in touch with us on a regular basis and responded to our questions with prompt Skype calls. Even though we’re teams that are thousands of miles apart, the communication has felt like we are working very close.

What was the team composition?

We met Ivan and he has stayed in touch with us from day one and continues to always be there to assist us. He has clearly listed the team assigned to our project and introduced us to other parts of the business that can help us with raising capital to further our project.

Can you share any outcomes from the project that demonstrate progress or success?

Our platform is still in the early days, but we have a very clear path on how to get to our final product. Knowing we have DRL working with us, we are confident we will deliver a platform that is high quality and provides value to our users. Our project already has potentials to grow and there’s no doubt the DRL team will be by our side for the whole journey.

How effective was the workflow between your team and theirs?

So far, we’ve been impressed with the fast responses, knowledge, proven experience and patience from the DRL team. They keep in touch with us regularly and are eager to help us figure out any hurdles we come across.

What did you find most impressive about this company?

Their knowledge in the technology field and speed of work on existing projects is what had us in awe. After our first meeting, we felt like coming up with more ideas to work on with DRL. We look forward to sharing more AI based ideas with Ivan and his team once we’ve past some of our major milestones for the SavStart project.

Are there any areas for improvement?

Please open an office in London, it’d be a joy to work even closer. Even though you do a superb job remotely.

Please describe your company and your position there.

I'm the CEO of Arcane Future. The company provides advanced analytical services for sites and applications. We help medium and small businesses build own analytical infrastructure and get insights from the data.

For what projects/services did your company hire DataRoot Labs?

Our client in a medical tourism industry requested the dashboards system to make the sales and marketing processes clear, predictable and efficient. Dashboards work faster when they connected to single datasource, so it was necessary to develop the data warehouse to collect data from CRM, CMS and Google Analytics systems.

What were your goals for this project?

Create the data warehouse on Amazon Redshift and together with the client's development team configure tracking from CRM, CMS to the DWH.

How did you select DataRoot Labs?

The project had time limits and we were looking for a company with the necessary expertise. Since I know that DataRoot Labs had already made similar projects we contacted and together developed a plan how make the DWH solution.

Describe the project in detail.

Once we started to work on this project we had 2 meetings with our client, his CTO and CBDM. On meetings we collected the requirements for the dashboards system and current product infrastructure. After this we together with DataRoot Labs made architecture of the DWH, and prepared technical tasks for the client's development team.

After tracking and DWH implementation DataRoot Labs tested the system and transfer to us all credentials and DWH documentation.

What was the team composition?

The team consisted of 2 data engineers and 1 project manager.

Can you share any outcomes from the project that demonstrate progress or success?

The speed of data loading from Redshift database was 2x-3x times faster compared to MySQL database.

How effective was the workflow between your team and theirs?

We communicated via Slack, they always respond in time. The workflow was managed via Jira board, to which we had access. We split the project on 1 week sprints and had a retrospective meeting every Friday.

What did you find most impressive about this company?

Due to their broad expertise in database architecture we made data warehouse which is convenient to use for an analytical work.

Are there any areas for improvement?

Overall we were satisfied with DataRoot Labs work.

Introduce your business and what you do there.

I’m the co-founder and COO of a startup organization. We work to improve clinical care for heart failure patients.

What challenge were you trying to address with DataRoot Labs?

We needed to develop a prediction algorithm.

What was the scope of their involvement?

DataRoots worked on one of our machine learning algorithms, which was a prediction algorithm. They used synthetic patient data and managed the machine learning process with that data to deliver an algorithm that would predict the timing of an adverse event. We had an API that made the algorithm usable as part of our technology stack.

They had a framework for machine learning projects, and we went through the entire process with them. Their team used a deep learning algorithm and TensorBoard as part of the testing and training components.

We finished the first algorithm. However, we need to build several algorithms. We similarly need a dataset to develop the next one. We’re now in discussions about what that dataset will look like.

What is the team composition?

Our main technical contact was Ivan (CTO) who was primarily involved in this project. He had a team member who supported him and joined our discussions too.

How did you come to work with DataRoot Labs?

We started our search with Clutch and considered a number of organizations based on reviews and recommendations from our network. My co-founder and I conducted interviews and evaluated several different companies. We narrowed our choices to DataRoot feeling that they would be the best fit for this type of project.

How much have you invested with them?

This project cost between $10,000–$15,000.

What is the status of this engagement?

We actually started this project around January 2020, and our partnership’s ongoing.

What evidence can you share that demonstrates the impact of the engagement? 

This was our first machine learning project with any organization. One of our goals was to learn the entire process. We had expertise from reading articles but not from actually working through the entire process. We learned a lot, as did DataRoot.

We got a high level of accuracy from the prediction algorithm. The accuracy of the algorithm was what we expected — 90% plus. From that perspective, it was a success. That part’s still a little unknown because that was all with synthetic data. We haven’t run the algorithm on real data, which we’ll explore with DataRoot.

How did DataRoot Labs perform from a project management standpoint?

It was a small project, and we primarily worked with one technical person. It wasn’t complex from a project management perspective.

This project ran during COVID-19, so we had to both adapt to what that meant for our organizations. The challenges that we experienced related to the pandemic. Communication around that adjustment could’ve been better. At times, it seemed that their team was slow to respond.

What did you find most impressive about them?

Their technical skillset around AI and machine learning most impressed us. They had deep expertise, which made us feel comfortable. Their team was up to date on the latest algorithms and able to guide us. They also spent a lot of time teaching and rationalizing their decisions for us.

Are there any areas they could improve?

It was a small project, so it didn’t have much of a process around it. However, it would’ve been helpful if we had a better understanding of the project’s timeline and process.

Do you have any advice for potential customers?

Go into the project with an open mind, especially if this is your first time working on an AI or machine learning project. There’s a lot to be learned. Make sure to come to the project with a good data set that you can explain. Short of that, it’s hard to get off the starting block.

Please describe your company and your position there.

I investigated several ideas of mine regarding augmented reality. Then I organised the project team and was investor/co-founder.

For what projects/services did your company hire DataRoot Labs?

I was in dire need of engineers who has machine learning skills (with computer vision experience). DataRoot Labs consulted me regarding technologies, provided staff with the necessary skills for this project and made technical leading of that engineers.

What were your goals for this project?

I wanted to have working prototype of the AI model. It is part of the interface based on augmented reality. It was important to use technologies that will work in a regular mobile browser.

How did you select DataRoot Labs?

I met the CEO of DataRoot Labs through a mutual friend. I know some other companies, but DataRoot Labs were more attractive to me regarding some points: we located in same city, speak same language and (the most important thing) they had much more experience of computer vision technologies than any other companies which I known.

Describe the project in detail.

I explained my idea (control of the game interface via the camera) to CTO of DataRoot Labs. We discussed about technologies, hardware, RND plan and other needs to develop what I wanted. We agreed to provide me engineers by outstaff model to develop working prototype.

What was the team composition?

There was two middle level engineers from DataRoot Labs, project manager from my team and me.

Can you share any outcomes from the project that demonstrate progress or success?

The first demo was appeared in first two weeks. And after 2 month team give working prototype. And then they worked on improvements. This prototype we used to test the concept/idea with focus group.

As a result received important information for making decisions about building the product base on that concept. In addition to this project, many other ideas about augmented reality were discussed and received several professional consultations

How effective was the workflow between your team and theirs?

The workflow was best as much as possible because all team could sit at one place, communicate easy, to see together on device each version. It's important because the tasks cause a lot of interactions with real testing device and real objects near device.

What did you find most impressive about this company?

Except the professionalism, easy going communication and pleasant staff, I was surprised regarding level of projects which this company did.

When you understand that the future is already here, because guys from DataRoot are already solve projects and tasks with machine learning that you could only dream about before.

Are there any areas for improvement?

I have no idea

Please describe your company and your position there.

I’m the founder and CEO of a consumer-facing business in the automotive aftermarket.

For what projects/services did your company hire data root labs?

We lacked internal data science. We wanted machine learning capabilities to develop the AI piece of the platform that would have a neural network predicting the probability of a given car failing.

What were your goals for this project?

We wanted the model built with a minimum of 0.85 accuracy and a ready API to connect with the rest of the platform.

How did you select this vendor?

I spoke with a few vendors based on their Clutch reviews. I picked data root labs for their deep understanding of the automotive field and data science, as well as their value proposition.

Describe the project and the services they provided in detail.

The aim of our project is to provide a service that predicts the probability of car a failing in the next 12 months based on the car’s previous history. They developed a machine learning AI model for our solution. Capabilities included normalizing data from millions of records and identifying and engineering the best variables for inference. They developed, tested, and refined an algorithm.

What was the team composition?

We worked with an architect and a data scientist. I communicated for the most part with Ivan (CTO and Co-Founder, data root labs), who was always ready to hop on the call and explain where we were at, what was done, and what was happening next.

Can you share any information that demonstrates the impact that this project has had on your business?

The model delivered by the team works fine and is well-integrated into our overall solution.

How was project management arranged and how effective was it?

They usually responded within a few hours of my emails and were proactive in addressing upcoming issues. Both of my contacts had an excellent understanding of what needed to be done. They communicated well and each had a commendable work ethic.

What did you find most impressive about this company?

For one thing, they exceeded my expectations by delivering the model with higher accuracy levels. Ivan’s data science team is particularly impressive. It is rare to find data scientists who understand the AI and machine learning field well, are able to articulate the concepts simply and are commercially aware enough to apply them successfully. I was impressed with their organization, their speed of communication, and their proactiveness.

Are there any areas for improvement?

No, I’m completely satisfied.

Introduce your business and what you do there.

I’m the CTO of the Ukrainian office of Dentsu Aegis Network. It’s a large multinational communications company that services about 400 companies and clients.

What challenge were you trying to address with DataRoot Labs?

We wanted a platform that could collect specific data from a variety of advertising sources. Then we wanted to implement AI technology that uses the aggregated data to evaluate performance, make future predictions, and generate reports for clients.

What was the scope of their involvement?

Over a 12-month period, DataRoot Labs developed an MVP version of the solution from the ground up. We’re currently working on future iterations. They set up our system on a computer cluster of various Kubernetes, PostgreSQL, and Redis databases. In its current state, the platform collects statistical data from sources like Facebook Ads and Google Ads, processes it, and generates reports with predictions regarding future performance, which helps our marketing team create better digital ads for our clients.

What is the team composition?

Max (Co-founder and CEO, DataRoot Labs) is our main point of contact, but we also work with a project manager. They assigned a core team of four resources that includes a solution architect, a UX/UI designer, and two data scientists. In the later stages of the initial build, another project manager joined the team. Now that we're working on future iterations, there’s also a frontend developer, backend developer, and a QA engineer.

How did you come to work with DataRoot Labs?

I asked my colleagues in the industry, and one of them recommended DataRoot Labs and spoke highly of their extensive expertise in the AI field.

How much have you invested with them?

So far, we’ve spent $70,000­–$80,000.

What is the status of this engagement?

We started working with them in October 2017, and the project is ongoing.

What evidence can you share that demonstrates the impact of the engagement?

I don’t have any concrete metrics that speak to their work, but our users have provided largely positive feedback. Based on our experience, I highly recommend the DataRoot Labs team.

How did DataRoot Labs perform from a project management standpoint?

We haven’t had any issues with their project management skills. We communicate via several channels—mainly Slack, Skype, and email. Our primary business communication tool is Microsoft Teams. Also, we schedule a weekly call to share updates and discuss our progress.

What did you find most impressive about them?

I really appreciate their prompt response time and their technical expertise.

Are there any areas they could improve?

At this point, we haven’t had any issues.

Introduce your business and what you do there.

BUFF.game is a blockchain-chased loyalty program that rewards users for playing games. I’m a co-founder and the CEO of the company.

What challenge were you trying to address with DataRoot Labs?

We engaged DataRoot Labs as a development partner. We needed to build our app’s architecture and infrastructure from scratch, and we wanted a tailormade solution to our needs.

What was the scope of their involvement?

We first provided DataRoot Labs with our vision, strategy, and business goals. Our teams then sat together and defined the product’s technical requirements. We created a design document that described the product’s architecture and planning. They then helped us put together an R&D team.

The development team built an MVP of the app over eight weeks. It features an AI that analyzes game behavior and determines the rewards that a player has earned. It also uses blockchain technology as its foundation. The dev team had to overcome challenges with scalability and security among other issues to achieve a working product. We conceptualized the application to overlay mobile and desktop games; however, for the time being, it only works for desktop. Another vendor built the UI/UX based on wireframes that DataRoot Labs and our team created. DataRoot Labs helped review the finished UI/UX to ensure it matched our technical specifications.

What is the team composition?

We worked with a system architect, a blockchain developer, and a frontend developer. Once we completed the MVP, the team grew to include a project manager, a machine learning engineer, one frontend and one backend developer, a QA engineer, a blockchain developer, and a deep learning engineer.

How did you come to work with DataRoot Labs?

We were looking for a partner who could meet our needs and had sizable experience in AI technology and blockchain. DataRoot Labs came with excellent recommendations from colleagues.

How much have you invested with them?

We spent between $250,000–$500,000.

What is the status of this engagement?

We started working with DataRoot Labs in May 2018, and the project is ongoing.

What evidence can you share that demonstrates the impact of the engagement?

DataRoot Labs met our goals and delivered high-quality products on time. The app has received positive feedback from users on its performance and how well it meets our requirements. Any bugs that arose were addressed within a day by DataRoot Labs. We’re highly satisfied with their work and their responsiveness.

How did DataRoot Labs perform from a project management standpoint?

The team has run the project smoothly so far. They understand our priorities on technology and business needs. We can scale accordingly to meet our delivery and timeframe goals and to respond to community needs. We primarily communicate through Slack and use Zoom for biweekly calls.

What did you find most impressive about them?

DataRoot Labs demonstrates in-depth expertise in AI and blockchain technology. They can address our business vision and work efficiently within our timeline and budget.

Are there any areas they could improve?

No, we’re highly satisfied with their work. I’ve been particularly impressed with their CTO and his in-depth knowledge and penchant for problem-solving.

Do you have any advice for future clients of theirs?

We found it helpful to give them our entire vision and our goals for the product and the business. They could understand our scope and deliver better work as a result.

Introduce your business and what you do there.

Promo.ua is a performance marketing agency. We do search engine optimization and PPC services on usability services, and we have a small Magento development division. I’m the CEO.

What challenge were you trying to address with DataRoot?

We needed to monitor the positions of multiple keywords across Google search results, which is usually on a per keyword basis. This was too costly for us, and so was keyword clustering. To find a cheaper process, we asked DataRoot to develop our clustering algorithms based on the data received from search engine results.

What was the scope of their involvement?

We initially requested this service, and it was supposed to cost much less per keyword than the service was charging us. DataRoot did research for a week and told us their per keyword cost would be 10 times less, so we went with them to provide this service. We had a few examples of what we needed, which we gave to DataRoot. They were good at figuring out the necessary functionality and had full involvement. After they developed the first part, we asked for a proposal for the keyword-clustering algorithm, and they came back with what we requested.

What is the team dynamic?

There are four people on our team.

How did you come to work with DataRoot?

They were recommended by one of my partners.

How much have you invested with them?

We have spent over $50,000 with them.

What is the status of this engagement?

We started working with them in April 2016, and the relationship is ongoing.

What evidence can you share that demonstrates the impact of the engagement?

We pay only 10% per keyword now, thanks to their work.

How did DataRoot perform from a project management standpoint?

They usually meet the deadlines. Sometimes they need more time, but that’s nothing major. For example, in one month, they may need a few extra days. One of my team members has been in contact with them more than I have, but as far as I'm aware, he hasn't had any problems with communication.

What did you find most impressive about them?

Their cost and quality ratio is impressive. I know a few people who develop projects like this, and nobody else would do it below $100,000.

Are there any areas they could improve?

We are happy with them. I hope they keep developing without raising their prices.

Introduce your business and what you do there.

I work for Bookimed, a medical tourism company. We provide solutions for treatment abroad for those who can’t get what they’re looking for. I am the Chief Business Development Officer of the company.

What challenge were you trying to address with DataRoot?

The business challenge was increasing our conversion rates. We have a staff of 25-30 doctor coordinators working for the company who have to deal with around 2,000 patients. When a patient comes to our website, they will leave an inquiry for getting treatment abroad which will then be processed by our team.

We found that conversions could be improved and have tried to work with DataRoot on finding the best possible match between patients and doctor coordinators. Doctors have different medical backgrounds and can work with specific kinds of people. The patients can also be different from person to person. 

What was the scope of their involvement?

In order for us to learn everything we can about the user, DataRoot developed an analytical data-gathering system that obtains behavioral data from our website. For example, potential customers will come from different countries, look at different pages, type in different queries and spend differing amounts of time on our pages.

We have more than 300 different behavioral parameters for the users on the website cultivating a neural network which will determine which doctor coordinator will match the specific needs of a customer. In order to find out what kinds of patients can work well with them, the neural network can analyze all background cases of doctor coordinators.

For this project, there were at least 2-3 people involved from DataRoot’s team. We are working on other projects with them as well.

As far as I know, they’ve implemented the data-gathering functionality of our website through user behavioral events which send information to our Mongo database. There is a neural network algorithm in place as well as a regression model for processing the data.

How did you come to work with DataRoot?

Our business owner is good friends with the owner of DataRoot. We decided to find a project which could prove interesting for both our companies. We’ve come to know the company and have looked at some of the other projects they were involved in. I knew that DataRoot would perform well.

How much have you invested with DataRoot?

The cost of this project was between $15,000 and $20,000.

What is the status of this engagement?

Over time, we’ve launched at least 10 projects with DataRoot. The current one started around 6 months ago. The product has been launched recently and is working well. We will probably enhance it with new data and user-behavioral events in the future. The website is constantly evolving. We can create new behavioral patterns on it by adding new types of pages and interactional user interfaces which can be counted as behavioral criteria.

Could you share any evidence that would demonstrate the productivity, quality of work, or the impact of the engagement?

DataRoot has setup a neural network which continues to learn from the patient-doctor interaction data. They have also implemented a sort of tutor through which doctors can teach the network about the items that are not interesting or appropriate for them. The system is working autonomously and doesn't require ongoing support from DataRoot’s side.

When we started the project, there was nothing in place that could influence our leads. The conversion rate has increased since this tool was implemented and it continues to grow. We can assign the appropriate patients to the appropriate doctors.

How did DataRoot perform from a project management standpoint?

We created multi-company groups on online tools like Slack, Skype and Workplace by Facebook. At least twice a month, we gather together with DataRoot’s team to brainstorm and coordinate our next steps. The collaboration has worked well from a project management standpoint with no problems that I can remember.

What did you find most impressive about DataRoot?

In spite of their specialists’ young age, they have a large amount of knowledge and skills. The main advantage of DataRoot is that their staff really loves machine learning, AI and data-mining.

Are there any areas DataRoot could improve?

I am not a specialist in artificial intelligence so I can't comment in this area. From an interpersonal perspective, the team is made up of great people.