- Works under real world conditions
- Handles adverse lighting
- Handles partial occlusion
- Handles complete in-plane rotation
- Handles small 3D Rotations
- and many more...
- Works with any form type
- Printed and handwritten forms
- Handles in-plane rotation
- Handles print quality variation
- 3D Rotation allowed for KYC
- and many more...
- Search using Semantic Similarity
- Image based similar designs
- Fast Query Times
- and many more...
- Mail us to know more
Applications include Analog Gauge Reader, 3D Localization, Object Recognition, Monocular Augmented Reality etc.
Over 7+ years experience in DL: both Convolutional and Probabilistic. Productized Form Reading, Aesthetic Image Scoring among others.
Developed solutions various problems in NLP: Text Classification, Topic Modelling, Document Search etc.
Experienced with developing multiple search solutions. Including custom ranking routines for Apache Lucene and adapting/extending queries for Apache Solr.
Developed and prototyped multiple solutions in Visual Localization (SLAM), Obstacle Detection, Obstacle Avoidance.
Reasons to choose us.
We have developed research products, end to end, from protyping to production systems. We have experience delivering on long term engagements as successful production systems.
We have over 15 years of research experience and over 7+ years of experience in productionizing research projects.
We allow for flexible IP terms, that guarantee success for our clients and while also keeping costs low.
We follow fast iterations using Agile like development cycles. We use best tools that allow our teams to be nimble and efficient.
Fastest answers to frequently asked questions
Do you only develop R&D prototype or do you develop full product?
Yes, we do full product development. To the extend that it contributes to core research product. If need be, we will work with your team to deliver on time.
We do not take up software development that is outside of the core product.
How long will the project take? or Can we do this project in 3 months?
We usually do a pilot with each customer that last anywhere from 4-6 weeks. Upon completion we will decide on longer term engagement, if we both decide that the problem is solvable, our expertise matches the particular problem and you as client, is happy with our work.
Usually we pick up problems that require deep technical expertise. Usually our customers come to us, when have either tried to develop it in-house or had an initial contractor with whom it was not successful. Due to the nature of such problems, they usually last anywhere from few months to multi-year projects.
We have xyz data, can you do Machine Learning for it and give me insights or develop a new product?
Our past experience has taught us that such exploration usually don't lead to business value. Thus such projects have premature termination or lead to unhappy middle managers.
We usually suggest customers to ask the reverse question, which is what is the problem that really bugs your customers. If you feel such problems requires Machine Learning, talk to us.
What value do you add when code/models are already available online?
Wonderful question: Such models help you develop your first iteration of solution or first proto as we call it. Soon you realize that such and such data points might not be available, or that you can't rely too much on your data quality. You might also realize that for very well cleaned data, those online solutions work well but your data is not clean. Or you might realize that your algorithms accuracy is 95% and you do 1 million data points per day and need need lesser than 10 errors per day. And unfortunately, your current system makes 50,000 errors per day.
This stage is usually where we come in. We help you narrow down and define the business problem such that the problem is solvable using current day tech, making viable and reasonable assumptions, using models that are general enough for the given problem. In short, when the online models have stopped helping you, that is when you call us.
How do I know if you are suitable for this project?
Talk to us. If we can't help you, we can connect you to our contacts either in Bangalore, India or San Francisco Bay Area, US.
Who owns the data?
Get in touch
To know more about our products and services
Please email us
#39, 8th Main Road