AI & ML solutions development
Zpoken AI is a machine learning lab which provides DS/ML and AI applications, services and development. Zpoken creates digital products for the business needs in various domains: computer vision, speech recognition & analytics, NLP products, recommendation systems, and Mathematical Models.
headquarters
other locations
-
33W 17th StreetNew York, NY 10010United States
Focus
Portfolio
Computer Vision for Robotics Repair: Wind Turbines
A major obstacle on the road to full automation is that the robot cannot find the edge of the blade without humans. Blade positioning is adjusted manually before each repairment session. Without being able to tell where exactly is the edge affected by oil and sand exposure, the robot cannot latch and start working independently.
Once on the edge, the robot uses geopositioning to figure out possible trajectories for the repair arm. Positioning calibration requires a reference point, and the current technology still requires human input to find that point. A self-sufficient robot would have to handle this by itself, too.
Solution Although the BR-8 robot has multiple cameras, we are employing the one embedded in the repair arm. The arm is the most dynamic part of BR-8, so using it to find the work area and a reference point for geopositioning calibration makes sense the most.
Arm camera is also used because it is a fisheye one. Such cameras provide a panoramic view (a great increase in robot’s safety) at the expense of distorting images. As a result, simply teaching the robot to process them is not enough: it does not see the way regular cameras or humans do. Image cropping with the use of machine learning mitigates this challenge.
To make the robot latch without human input, we taught it how to measure the distance between itself and the edge of the turbine blade. Classic computer vision methods helped to achieve a near-perfect distance prediction; the little deviation has no impact on BR-8’s work and is not a source of danger for the robot. CV also enables BR-8 to measure the angle between the robot and the blade edge, so manual repositioning of blades is not required for repairs.
For geopositioning calibration purposes, we are working with an arm attachment that has four dispensers on it. Our algorithm finds a particular dispenser and transmits its placement so that a proprietary Rope Robotics solution can learn the new geolocation. This task is impossible to solve with only computer vision. We are employing a lightweight neural network to segment fisheye images and “show” the robot where to look for the center of the reference dispenser.
Results
The autonomous robot latching solution is complete, and it does the job both indoor for testing purposes and outdoor for real-life use. Geopositioning calibration is still a work in progress: the current version can be improved upon before moving into production. A later iteration would function with just one dispenser on the arm attachment instead of four.
Both solutions work on Python, which makes them light enough to install on a Raspberry Pi. Using a low power consumption implementation is vital for outdoor work, as wasting time on robot recharging increases the downtime for wind turbines.
As turbine manufacturers get to lower their warranty expenses with the help of Rope Robotics, they may get more competitive about the pricing. With alarming reports on climate change, easier access to alternative energy is welcome more than ever.
Technology: Computer Vision, Machine Learning, Neural Network
Reviews
the project
AI Voice Bot Dev for Contact Center
"They looked at the project from very different angles and provided very good consulting."
the reviewer
the review
The client submitted this review online.
Please describe your company and your position there.
CallWay is contact center service provider with custom software for different market needs. My position in the company is product lead.
For what projects/services did your company hire Zpoken, and what were your goals?
We had a challenging project to replace our contact center operators by AI voice bot. The goal was to decrease the calls traffic on our operators and cover more calls during the high hours.
How did you select Zpoken and what were the deciding factors?
We have only several providers on our market who can deliver such services. The main factor for us to make a decision was that they have enough experience and courage to develop challenging solutions in the AI field
Describe the scope of work in detail, including the project steps, key deliverables, and technologies used.
Conversational voice bot integrated to the contact center ip telephony using MRCP protocol. Speech recognition and generation was made by the Zpoken solution using machine learning and python coding. The deliverables was to receive call and correctly recognize a ride request.
How many people from the vendor's team worked with you, and what were their positions?
Their was 3 people from the Zpoken side: project manager, machine learning engineer and integration specialist
Can you share any measurable outcomes of the project or general feedback about the deliverables?
We have decreased the call traffic from the operators using AI voice bot to proceed the taxi requests by 30%
Describe their project management style, including communication tools and timelines.
They use email and messenger chats to faster the communication, also they use git, slack and notion for planning, code delivery and discussions
What did you find most impressive or unique about this company?
They have experience in different tech fields from machine learning and classical software development to hardware installations and high load applications development and support. So they looked at the project from very different angles and provided very good consulting before the project start
Are there any areas for improvement or something they could have done differently?
No, we have transferred R&D project to them with goal to try different approach in contact center, so they done good job on this challenge
the project
AI Dev Staff Augmentation for Farming Startup
"We have improved our product with several new features."
the reviewer
the review
The client submitted this review online.
Please describe your company and your position there.
I am a Chief Scientist at a Swiss startup aspiring to help farmers globally to improve the efficiency and sustainability of their production using a unique constellation of cutting edge technologies including hyperspectral imaging and artificial intelligence.
For what projects/services did your company hire Zpoken?
We needed to double our AI team in a short period of time to run 5 new research projects at the same time and make the product pipeline once we got the results.
What were your goals for this project?
The goals of the project included the development of a host of technologies to significantly improve the quality of the available satellite imagery to facilitate the commercial deployment of Gamaya digital agronomy services.
How did you select this vendor?
During the course of extensive selection process I have engaged and worked with a number of external teams on a number of experimental projects. I have rapidly become aware that the Zpoken team provided he optimum combination of technical skills, experience and project management capabilities. Consequently, Zpoken become the team of choice for the majority of company's outsourcing needs.
Describe the project in detail.
The collaboration started with an experimental project, were the project description and the expected results were provided by me. The Zpoken team have been able to rapidly internalise the project objectives, define and implement a development roadmap and deliver tangible results within the defined timeline. We have subsequently extended the original project, as well as the scope of collaboration to multiple parallel streams of activity. The majority of these initiatives were successfully delivered.
What was the team composition?
The team was constituted by an experienced project manager as well as a team of 4 exceptionally skilled data scientists and software engineers. The project further included web-development component that facilitated the demonstration and visualisation of the obtained results. A total of six people were involved in the project from the Zpoken side.
Can you share any outcomes from the project that demonstrate progress or success?
We have improved our product with several new features that give our customers more options of agriculture analytics.
How effective was the workflow between your team and theirs?
Machine learning engineers were in the direct management by Gamaya, but PM from Zpoken AI side helped a lot to organise processes and manage tasks.
What did you find most impressive about this company?
We found it impressive that each time share a concern or edits, they immediately deal with it and get back with the solution. Zpoken AI also provides discounts for the GPU cloud servers depending on the project needs. Which makes possible to train neural networks in a very fast mode. We got up to a 100% discount on cloud GPU instances on some projects.
Are there any areas for improvement?
Communication and development was good enough. Preparation stage is really important in every AI project, so you need to spend the proper amount of time in pre-production understanding the goals that you can achieve with Machine learning/Deep learning/Data science.
the project
AI Dev for Machine Learning, Data Science & AI Company
"They’re conscious of how they can grow and improve because they take a professional approach to collaboration."
the reviewer
the review
A Clutch analyst personally interviewed this client over the phone. Below is an edited transcript.
Introduce your business and what you do there.
I’m the CEO and founder of 20 Thousand Leagues.
What challenge were you trying to address with Zpoken?
For a client of ours, we worked with them on computer vision machine learning for the purpose of improving a maintenance solution.
What was the scope of their involvement?
We’re building a renewable energy solution using robotic process automation for our client.
They were getting involved in building out a new network model to help handle the different types of object recognition, detection, and 3D reconstruction.
The project is still ongoing, but we’re working to make the operation stable in a consumer environment. That way, we’ll be able to run maintenance on the renewable energy solution’s maintenance.
What is the team composition?
I work with 3–4 people. Michael (Co-Founder & CBDO) and Anton (CEO) are my main points of contact.
How did you come to work with Zpoken?
I know them through my personal network, and they’ve worked on other projects with me in the past.
How much have you invested with them?
We spent about $45,000 on the project.
What is the status of this engagement?
They joined the project in May 2019, and our last phase was completed in March 2020. We’re planning on moving forward with them in the future.
What evidence can you share that demonstrates the impact of the engagement?
Our client was very happy with the work. In fact, they’re continuing our partnership based on the success of our past projects together. They were able to take us from the initial research phase into operations. By winning a new contract, succeeding in a technical proof of concept, and converting the account into an operation solution, they proved themselves to be very helpful to us.
How did Zpoken perform from a project management standpoint?
They’ve been excellent with communication all the way through.
What did you find most impressive about them?
They’ve gone above and beyond. When things haven’t gone as planned, they’ve redone work and covered the cost of doing so. They’re conscious of how they can grow and improve because they take a professional approach to collaboration.
the project
Custom Software Dev for Business & Financial News Platform
"They had an impressive quality of work and ability to complete projects quickly."
the reviewer
the review
The client submitted this review online.
Please describe your company and your position there.
Premium, global financial news service. CEO.
For what projects/services did your company hire Zpoken, and what were your goals?
Low latency natural language processing of unstructured regulatory filings.
How did you select Zpoken and what were the deciding factors?
Background, price and experience.
Describe the scope of work in detail, including the project steps, key deliverables, and technologies used.
Taking unstructured text and pulling out relevant data to create information for our newsroom.
How many people from the vendor's team worked with you, and what were their positions?
2, project manager and developer.
Can you share any measurable outcomes of the project or general feedback about the deliverables?
Very good results. Exceeded expectations.
Describe their project management style, including communication tools and timelines.
Regular, timely and clear communications throughout the project.
What did you find most impressive or unique about this company?
They had an impressive quality of work and ability to complete projects quickly.
Are there any areas for improvement or something they could have done differently?
No.
Zpoken has the experience and technical knowledge on software development and machine learning. Through this, the team was able to provide good consultation before starting the project. They maintained email and chat communications throughout for fast correspondence.