Who are we?

We are a team within Optisol Business Solutions, specializing in Data Analytics and Machine Learning. We think machine learning will revolutionize that way we use technology to solve everyday problems. It will help us automate mundane things so that humans can work on more important tasks. Automation is the end goal of AI. We try to equip ourselves with the knowledge of the latest happenings in this field so as to help our clients reap the benefits.

What do we do?

We specialize in building Machine Learning Models in Computer Vision, Conversational AI and Process Automation.

  1. In Computer Vision, we build light weight models that can run on a small, cheap hardware to perform object detection, image classification and activity recognition. This is useful in areas like Industrial Safety, Security etc.,
  2. In Conversational AI, we build models that can parse and interpret free text and answer simple questions from the text. In technical jargon, we build automatic intent generation models for chat bots.
  3. In Process Automation, we build RPA (Robotic Process Automation) robots using UIPath to automate repetitive tasks. For ex: automated invoice processing. In this realm, we bring our expertise in AI to build cognitive robots that can perform higher order tasks like parsing and interpreting emails and instruct the robot what to do.

What do you mean by Data Analytics?

Data Analytics is a broad term that includes all kinds of work that can be done on data. For example: we can analyse data to see understand something new about the business that generated the data. This is called Business Intelligence. We can use the insight we gain from analyzing data to make predictions about the future. This is called Predictive analytics. This helps stakeholders in business make better decisions. We provide both Business Intelligence solutions and predictive analytics solutions to our clients.

What is Machine Learning?

Machine Learning is training a software algorithm to make decisions based on past events. It is similar to how humans learn from our past events and mistakes. These software algorithms are fed some data, for example: a video feed from a CCTV camera with humans in the video feed marked or identified. From this the algorithm learns to detect humans in any video feed. Another examples could be feeding the algorithm all the archived help desk chat data from human operators assisting customers in the past. The algorithm will then learn to answer similar help desk questions of customers

Do you have work samples I can see?

Yes! Please check out selective sample in our ‘samples‘ page. Our exhaustive list of videos samples showing all our work is published at this  YouTube channel

What kind of data do you work with?

We work with standard tabular data that is in a spreadsheet or a database. We also work with unstructured data like text from website(s), twitter feeds etc., Images and videos are another important data format that we work a lot with. We can also work with any other domain specific data formats that we may haven’t even seen before. We have been working with different kinds of data for a while. So any data format is a fair game for us

Wow! sounds fancy? How much does it cost to train a Machine?

Well! it depends. But usually no more than developing a mobile app or website. The type of data and its complexity usually determines the cost. Why not contact us to see how much it costs.

Tell me about your past projects

Some of our past and current projects include

  1. Detecting Personal Protective Equipment compliance in a shop floor for CCTV camera feed. Our models take the video feed and automatically and continuously detect if workers are wearing safety vests, harness, hard hats etc., and notify violations to Supervisors to prevent injuries and law suits
  2. Detecting cell phone usage where it is prohibited, also from CCTV camera feed
  3. Detecting humans in a vehicle path (a fork-lift or other vehicles from mounted cameras on the vehicles) and trigger speed control systems
  4. Building smart chatbots that can learn by consuming vast quantity of data from websites like investopedia etc., and answer any customer query in that domain
  5. Building smart automation solutions that can automate typical back-office processes saving time and reduce errors.

Do check out our youtube channel where we have posted some of these Machine Learning models in action

Do I need Machine Learning in my app or site?

You need Machine Learning if you say yes to any of the following

  1. Do you offer recommendations to to your users?
  2. Do you personalize the user’s look and feel?
  3. Do you aggregate data from the web and show to your users?
  4. Do you offer chat or live support to your users?
  5. Do you or plan to offer speech recognition to your users for more natural engagement?
  6. Do you offer services like summarization, tagging of content?
  7. Do you or plan to offer face or image search, face recognition and object detection?
  8. Do you just want to have a cool product?

What are these BigData, Cloud, AI?

These are just some buzz words we technologist throw around to impress people. If you have pictures in google or apple cloud you already use big data. If you have gmail or similar email you are already using the Cloud. If you use Youtube or Netflix you are already using AI. So does your users. Don’t let the buzz words confuse you. Talk to us. Let us help you build the right mix of these buzz word technologies work for you.