DLabs.AI

BACKGROUND

Please describe your company and your position there.

Chief Brand Strategy and Community Partner. Our company simplifies international education via connecting students, universities, and a community all in one platform.

OPPORTUNITY / CHALLENGE

For what projects/services did your company hire DLabs.AI, and what were your goals?

We hired DLabs.AI to help us with extending our platform’s web and mobile app functionalities.

First of all, the main challenges were creating a new search engine, improving the data quality, and creating an ML solution helping with cost optimizations relating to the platform’s cloud service.

Additionally, we wanted to provide the best experience for our users by creating an ML recommendation system matching users to majors and courses.

In the beginning, the DLabs.AI team was responsible for delivering a Machine Learning solution. But when we noticed their vast knowledge and experience in web and mobile development, we decided to entrust them with building a mobile and web application.

SOLUTION

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

We found Dlabs.AI on Clutch. The main reason for the cooperation was the extensive experience of the DLabs.AI team in creating Machine Learning projects.

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

DLabs.AI supported us with the development part and organized their developers' pipelines. The Delivery Manager has been working closely with us to understand our business needs best and create solutions to our expectations.

Additionally, we had regular meetings with the DLabs.AI team to discuss the work status and talk about the next steps. They started with a detailed overview and analysis of every step of their work. DLabs.AI team always emphasized data quality and choosing the best technology to give our users the best experience.

To create this project, the DLabs.AI team used technologies such as Python, React, React Native, TypeScript, FastAPI, AWS, locust, pytest, PyTorch, Docker, ElasticSearch

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

DLabs.AI provided skilled, experienced experts to deliver an end-to-end AI solution. We worked with 4 Frontend Developers, 2 Backend Developers, 2 Data Scientists, Project Manager, Quality Assurance/Tester, and Business Analyst.

RESULTS & FEEDBACK

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

The solutions provided by DLabs.AI led to cost optimizations of around 38%. Moreover, their implementing auto-scaling cut the number of tasks running in quieter parts of the day by 75%

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

The DLabs.AI team worked in the Scrum methodology. The Delivery Manager carefully planned the team's work and made sure that everyone involved in the project knew what the most important business goals and priorities were in a given sprint.

To make sure everyone in the team knew what stage the work was at, we used Jira to manage tasks. To communicate efficiently, we used Slack, which was very convenient and allowed us to contact each other immediately.

We discussed the most important plans for the near future and discussed the best solutions during regular meetings. These meetings were crucial for us, and we learned a lot. The DLabs.AI team was very professional. They always tried to advise on the best solutions and be flexible to meet all our business needs.

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

Above all, we were impressed by the professionalism and great enthusiasm of the DLabs.AI team. They are passionate about what they do and try to do it as well as they can. They are always open to new ideas and share their suggestions.

The most impressive part of the project was the implementation of an utterly self-developed algorithm based on neural networks, which creates representations of students and courses and recommends the best fitting courses. The DLabs.AI team introduced an explainability module that indicates why a course was recommended.

The implementation of such a tool gives additional capabilities to our employees, reduces their bias, and expands their knowledge of the real needs of users. It has become an essential part of how our company operates, allowing us to get to know our users better and understand the technology behind these innovative solutions.

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

Nothing significant to add.

BACKGROUND

Please describe your company and your position there.

Da Vinci Studio is a programming company, offering dedicated web and mobile applications, solutions using machine learning, artificial intelligence and developer outsourcing. I’m CEO at Da Vinci Studio.

OPPORTUNITY / CHALLENGE

For what projects/services did your company hire DLabs.AI?

We hired DLabs.AI specialists for a research and development project related to the digitization of processes and optimization with the use of artificial intelligence.

What were your goals for this project?

The goals of our cooperation were:

• build a roadmap for the development of a machine learning component of our system
• develop an AI-based algorithm for extracting information from emails that will meet certain accuracy

SOLUTION

How did you select DLabs.AI?

We selected DLabs.AI due to their experience in similar projects. DLabs.AI delivered us a valuable offer at a good price and paid attention to our needs.

Describe the project in detail.

The aim of the project was to create an algorithm for content analysis and extraction of key information from e-mails. That information is weight, height, width, depth, the volume of the package, number of packages, company, container type, etc.

The scope of work included:

• Analysis of requests for a quotation
• Preparation of input data for the algorithm based on the analyzed requests for a quotation
• Preparation of an extraction algorithm
• Training of the neural network aimed at data extraction
• Neural network improvement.

Key deliverables: An algorithm that can predict message intent with 85% accuracy and can extract key information from e-mail with 63% F1-score.

Technology used: Python, Google Cloud AI Platform, Tensorflow, PyTorch, Pandas, NumPy, Matplotlib, Scikit-learn

What was the team composition?

Mainly three dedicated specialists - ML Engineer, Data Scientist and Project Manager

RESULTS & FEEDBACK

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

We are working with DLabs.AI since June and they showed us that they can be a trusted technological partner. The team’s design work aligns with our vision and we’re delighted with that.

How effective was the workflow between your team and theirs?

The workflow was highly effective. They’re always available for video-calls, e-mails and chat on Slack.

What did you find most impressive about this company?

They had a sense of ownership over their work and always informed us on what is going on our project. I think this is the point.

Are there any areas for improvement?

None. I don’t have anything bad to say about DLabs.AI, because our cooperation goes really well.

BACKGROUND

Introduce your business and what you do there.

I’m the CEO and founder of AdEmotion, which is a digital marketing agency based in Poland.

OPPORTUNITY / CHALLENGE

What challenge were you trying to address with DLabs.AI?

We needed a partner to build a custom AI platform. 

SOLUTION

What was the scope of their involvement?

DLabs.AI developed the frontend and backend of the platform. Their team created our algorithm’s architecture too. Our team gathered data that assisted that process.

The platform sells special digital campaigns on Facebook to users. The campaigns feature dedicated ads, graphics, and copy target to people according to their personality.

What is the team composition?

We interacted with about eight people in total. However, not everyone was involved in the project on a regular basis. Their team included two data scientists, four developers, a project manager, and the company’s partners. 

How did you come to work with DLabs.AI?

A colleague referred us to them. We received a few proposals but chose DLabs.AI based on their background and engagement. There was also chemistry between our teams.

How much have you invested with them?

We spent more than €200,000 euros (approximately $238,000 USD) on the project.

What is the status of this engagement?

We worked on the project between April 2017–December 2018.

RESULTS & FEEDBACK

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

We’re satisfied with the results. DLabs.AI created a high-quality product without any bugs. It also achieves good market results. We contracted Dlabs.AI again to build another platform. 

How did DLabs.AI perform from a project management standpoint?

We communicated via calls and had weekly meetings to review sprints and progress reports. Their team worked within our budget and timeline. They were also available and responsive. Their project manager performed well.

What did you find most impressive about them?

They were professional in their project management and communication. That’s important to me. 

Are there any areas they could improve?

No, I can’t think of ways for them to improve.

BACKGROUND

Please describe your company and your position there.

We add physical items inside game and expand ways how players experience the game. For this startup I have a CTO role and it means overlooking every tech related process

OPPORTUNITY / CHALLENGE

For what projects/services did your company hire DLabs Sp. z o.o.?

We were in tight runway finding product-market fit and in the same time developing product MVP version to test out our assumptions about potential customers.

We were cooperating with DLabs Sp. z.o.o. developing machine learning algorithm as well as web-based customer onboarding site

What were your goals for this project?

The main goal for these projects were - build potential solutions as fast as possible and test product-market fit

SOLUTION

How did you select this vendor?

They were recommended and introduced to us by our investor.

Describe the project and the services they provided in detail.

The machine learning solutions goal: develop an algorithm as a microservice that analyses whether a player inside the game can be monetized.

The second projects goal: develop frontend web application and backend API solution to automate customer onboarding process

What was the team composition?

The projects teams had about four people, but certain developers would come and go depending on the skills needed. We did have a stable project manager who was our direct and permanent point of contact.

RESULTS & FEEDBACK

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

Our solution was developed and launched as initially planned. There were some delays, that were caused by a lack of real data that would allow to test algorithm. Those delays were caused by stagnatng onboarding process

How was project management arranged and how effective was it?

Project management was very good. Team regularly complied with deadlines and always communicated potential delays . We communicated exclusively with Slack, Jira, and Google Meet.

What did you find most impressive about this company?

I appreciate that we have direct access to the project managers and the development team. Technical things can easily be lost in translation. They prevent this issue by keeping communication channels open.

Are there any areas for improvement?

There is not that much they could improve. My suggestion for the tech would be to document setup instructions as much as possible, documentation was great, but sometimes small bits of information were lost.

BACKGROUND

Please describe your company and your position there.

I am the Director of TV CONTROL, a company that specialised in Content Metadata management and delivery. We collect, enhance and delivery daily TV listings information to broadcasters and pay TV providers. This is done via a custom build CMS platform that is developed from scratch in order to optimise authoring and automate data entry operation.

OPPORTUNITY / CHALLENGE

For what projects/services did your company hire DLabs Sp. z o.o.?

We are using Dlabs services for a project that requires data analysis, deep search and syntax/grammar pattern recognition.

What were your goals for this project?

We aim to build an engine that will take all the laborious tasks of human data entry and attempt to create an AI data construction engine that will attempt to replicate the creative authoring process of editors.

SOLUTION

How did you select this vendor?

We reviewed various companies offering Machine Learning and we valued the Dlabs experience in past projects and their approach to our needs.

Describe the project in detail.

Text construction is a new field in ML and not yet fully developed in order to replace human creativity. Nevertheless, specific tasks in the authoring process can be analysed, and become templated automatic actions performed by a machine. Overall time can be saved from end 2 end operations which is the goal of this project.

What was the team composition?

A senior Data scientist, a project manager and later a software developer, all form Dlabs,

RESULTS & FEEDBACK

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

Project is still ongoing...

How effective was the workflow between your team and theirs?

So far the communications are good and effective.

What did you find most impressive about this company?

They have great attention to detail and background knowledge in the subject.

Are there any areas for improvement?

I it too early to have an opinion as the project is ongoing.

BACKGROUND

Please describe your company and your position there.

We are SaaS company in marketing space. We have two main products we offer: Unamo SEO and Unamo Social Media.

OPPORTUNITY / CHALLENGE

For what projects/services did your company hire DLabs Sp. z o.o.?

We need a company that we could partner with to create new features for our product. We needed a strong team with Machine Learning skills.

What were your goals for this project?

The goal of the project was to provide software to detect logo appearance on images and photos using advanced ML algorithms.

SOLUTION

How did you select this vendor?

It was recommended by one of our investors, we checked couple of other vendors, but communication and involvement of DLabs team helped us to make that decision.

Describe the project and the services they provided in detail.

They found a way to achieve goals we needed to reach, helped us to create experiments to confirm the idea and then provided development services to implement whole solution.

What was the team composition?

The project team was about 3-4 people but it was changing over time depending on the phase of the project.

RESULTS & FEEDBACK

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

Thanks to DLabs we were able to deliver new features much more faster than we would do that in house.

How was project management arranged and how effective was it?

The team joined our scrum meeting and quickly worked like a one team. We were using JIRA, GitHub and Slack.

What did you find most impressive about this company?

They exhibit great knowledge about machine learning.

Are there any areas for improvement?

Sometimes we felt that there is not enough communication during sprints, but every time we reached them we received detailed information about where we are.

BACKGROUND

Please describe your company and your position there.

The company is a developer of energy management software, which, among other things, supports rational resource management. My position is Director of software development

OPPORTUNITY / CHALLENGE

For what projects/services did your company hire DLabs Sp. z o.o.?

We hired them to create modules for process monitoring.

What were your goals for this project?

After the proof-of-concept phase, a microservice was developed. It makes it possible to use all of the functionalities using REST API and is currently integrated with our system.

SOLUTION

How did you select this vendor?

There were five other possible vendors, but most of them were disqualified due to inexperience.

Describe the project and the services they provided in detail.

DLabs started with our business needs. They performed data preprocessing and exploratory analysis. They considered several possible solutions and chose those that were the best fit for our business needs and then implemented them.

In this project, DLabs sp. z. o. o. researched and then developed two new modules for process monitoring. One of them was designed for internal use by the client and the other will be offered to their customers. The first module was responsible for signal prediction and anomaly detection. The module can track multiple signals, such as electric power or water usage. The second module is used to predict faults in the server room.

What was the team composition?

Two data scientists were involved from the beginning and a machine learning engineer joined the project later.

RESULTS & FEEDBACK

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

It was a real pleasure working with DLabs. We would gladly work with them again without hesitation. The DLabs team was very experienced and had a great mastery of machine learning algorithms. They took the time to listen and understand the company’s needs and issues. Their team did an excellent job of dedicating time and focus on each step of the process, which helped when delivering great products.

How was project management arranged and how effective was it?

They communicated with us through email and video calls. DLabs was flexible in terms of project requirements and the project deadlines were respected.

What did you find most impressive about this company?

Their management team was very involved and responsive. They followed best practices for software development and used modern technologies.

Are there any areas for improvement?

I can't think of anything.

BACKGROUND

Introduce your business and what you do there.

I'm a data scientist at DrivenData, an organization that runs predictive modeling competitions in which data scientists from around the world compete on data problems with a social impact.

OPPORTUNITY / CHALLENGE

What challenge were you trying to address with DLabs?

We reached out to DLabs for data cleaning and preparation, exploratory data analysis, machine learning, and content development services helping us get several competitions from just an idea to being fully prepped and ready to launch.

SOLUTION

What was the scope of their involvement?

DLabs's data scientists helped us get ready for two competitions, taking our loosely sketched ideas and a data set. They cleaned and prepared the data, did some exploratory analysis, framed a predictive modeling question, and then benchmarked the performance of several different machine learning approaches. They also developed the participant-focused content,  providing context and motivation as well as setting up the technical problem statement and explanation of data and performance metric. Components of the work included writing content, machine learning, and visualization in order to make sure they could convey to our team what the expected performance was for learners and confirm that the problem was difficult but feasible.

There was some ambiguity in the beginning, so they had to figure out and choose a more specific direction to take with the predictive challenge. 

What is the team composition? 

We primarily work with a principal and a data scientist on their team.

How did you come to work with DLabs?

I'd been in communication with their founding team for a while before reaching out to them for this project. They've worked on other projects that use similar technical stacks, so they were a natural fit. 

How much have you invested with them?

The first project was about $3,000, the second was $5,000, and the current project we're working on is $20,000.

What is the status of this engagement?

We started working on the most recent project in February 2018 and completed it within a month. We're now working with them on a longer software development project that should take about six months. 

RESULTS & FEEDBACK

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

We've been very happy with the quality of their work. We've had an unusually low level of support requests for each project they've worked on, meaning people are having fewer issues understanding what they're supposed to do because the data set was properly prepared and the content was extremely clear. 

How did DLabs perform from a project management standpoint?

They're great at flagging questions for us, reaching out whenever they need clarity on something and bringing up key issues during our check-ins. They primarily communicate with us through email, video calls, Slack, and GitHub.

What did you find most impressive about them?

They're extremely responsive and their management team is very involved. They also invest a lot in modern technology and software development best practices, so they continue to be great to collaborate with. 

Are there any areas they could improve?

I can't think of anything.

Do you have any advice for potential customers?

They have a talented software engineers and also experienced data scientists on their team, which is pretty rare. It especially makes sense to work with them on projects that could benefit from involves both skillsets. 

BACKGROUND

Please describe your company and your position there.

I'm an associate at Alfabeat, an early-stage venture capital firm that’s investing in B2B software companies. I’m responsible for launching AI projects with our portfolio companies.

OPPORTUNITY / CHALLENGE

For what projects/services did your company hire DLabs Sp. z o.o.?

We hired them to develop custom AI algorithms for our portfolio company.

What were your goals for this project?

Our goal was to recognize patterns on real images. Since the data pool was not big enough for AI image analysis, another algorithm had to generate synthetic data first.

SOLUTION

How did you select this vendor?

We hired DLabs Sp. z o.o. because of their experience in working with synthetic data and image analysis. Since their office was close to ours, cooperation would be easier.

Describe the project and the services they provided in detail.

DLabs Sp. z o.o. developed a set of custom algorithms. The first algorithm generated synthetic photos based on an image sent, while the second was trained to recognize patterns on the photos based on the synthetic pool of data.

What was the team composition?

We worked with machine learning engineers, data scientists, and 1–2 developers.

RESULTS & FEEDBACK

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

Their work gained a competitive edge for our portfolio company.

How was project management arranged and how effective was it?

It was effective. Our communication was good.

What did you find most impressive about this company?

They deliver what they say that they will deliver.

Are there any areas for improvement?

I was satisfied with the cooperation.

BACKGROUND

Please describe your company and your position there.

Our business offers enhanced social listening and media monitoring tools. One of our tools aggregates various documents and makes it possible to analyze them.

OPPORTUNITY / CHALLENGE

For what projects/services did your company hire DLabs Sp. z o.o.?

We reached out to DLabs to help us determine if a statistically significant relationship exists between user engagement expressed as a document’s popularity on social media and the semantic and syntactic features derived from that document.

What were your goals for this project?

Our goal was to have DLabs's data scientists help us prepare a text corpus including over 600 thousand documents written in Polish. The documents were from several different sources and included metadata about popularity.

SOLUTION

How did you select this vendor?

We selected them based on recommendations.

Describe the project in detail.

They derived statistical, semantic, and syntactic features and performed an analysis of the statistical relationship between these features and different popularity descriptors available in the document’s metadata. Additionally, they used advanced machine learning techniques to train models to classify popular documents.

What was the team composition?

We worked with a data scientist, a machine learning engineer, and a backend engineer.

RESULTS & FEEDBACK

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

DLabs obtained very promising results.

How effective was the workflow between your team and theirs?

The workflow was smooth and close to perfection.

What did you find most impressive about this company?

They took a professional approach to cooperation and provided great feedback.

Are there any areas for improvement?

The partnership was so great that I can’t expect any improvement.