We do AI/ML, IoT, computer graphics
DeepCraft is a research and development company providing state of the art solutions for customers at any phase of business or software development — from startups at the ideation stage to mid-sized businesses and large enterprises with a focus on optimization and expansion no matter the industry they operate in. Our purpose is to benefit our clients — their ideas, products, and services.
Our core technology strengths are artificial-intelligence/machine-learning, data engineering, IoT. We have extensive experience across a wide range of industries, including the public safety, medical and enterprise resource planning sectors.
Focus
Portfolio
High throughput radar system with AI
DeepCraft had created software and AI/ML solutionfor a next generation security system which covertly deploys as a layered multi-sensor platform to detect and combat active weapons and explosives threats before they occur. The fully integrated platform cost-effectively shares proprietary networked machine learning software for maximum effectiveness to covertly secure perimeters (i.e. parking lots, walkways), access/egress points, and facility interiors.
Acoustic echo cancellation
An active acoustic cancellation module that eliminates echoes during calls providing clean audio conversation.
Email task extractor Bot with AI
A smart system that could extract tasks from email and push into a EPR system for further management. The module could handle quite versatile emails and styles. It could extract assignees and deadlines if available.
Resume processing Bot with AI
Resume processing module could process thousands of customly formatted resumes in seconds and extract all necessary information from them, organize them into common structured data format for furtner matching with job requirements.
High performance AI model inference library with for a noise cancellation system
Modern ML libraries use Python as programming environment and for inference mode they rely on GPU acceleration. Still there are platforms which does not have hardware acceleration and should rely on CPU optimized inference library. Our solution was recorded to be up to 13 times fastster than tensorflow on CPU.
Reviews
the project
Video Segmentation C++ Library Dev for Audio Quality Startup
"The quality of the code and techniques used for speed optimization were very impressive."
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 president and co-founder of Krisp, an audio quality startup.
What challenge were you trying to address with DeepCraft?
We developed a machine learning model with transit flow. We had the challenge of developing an SDK in C++ so that it could run on different platforms and needed help.
What was the scope of their involvement?
They helped with porting our Python code into a C++ library for video segmentation. The solution removes the background of a person and replaces it with any photo you have.
What is the team composition?
Vardgas (CTO & Co-Founder) lead the project. There were also two software engineers.
How did you come to work with DeepCraft?
I knew that they had a very skilled workforce in C++. I knew they had experience in machine learning models. They had the combination of skills I was looking for. I met Vardgas at a meetup and he presented his company. We started with a small project and then decided to outsource the whole project to them.
How much have you invested with them?
We spent about $40,000.
What is the status of this engagement?
We worked together from December 2019–April 2020.
What evidence can you share that demonstrates the impact of the engagement?
They’re really great. They’re like an extension of our team. They provided the SDK with unit tests and a very nice interface. We continue to use the product they delivered. It’s been a good experience.
How did DeepCraft perform from a project management standpoint?
Every two weeks, we’re presented with the breakdown. The process was very straightforward. We added their team on our Slack. We had very direct communication. It was a very transparent and effective process.
What did you find most impressive about them?
The time management was really good. They delivered everything as promised in the timeframes set. The quality of the code and techniques used for speed optimization were very impressive.
Are there any areas they could improve?
They could have deeper knowledge in machine learning and network architecture which would give them more advantages over other outsourcing companies.
Do you have any advice for potential customers?
Define the scope of the work and be sure about the technologies that would be used for the project. We knew what we wanted and what technology should be used for building the SDK and the scope of work was defined correctly.
The DeepCraft team acts as an extension of the internal team. The process was straightforward, the quality of work was high, and the client continues to use the product that DeepCraft developed.