Neural Spark

We are in the era of neuro-cognitive machine intelligence where robotics and AI come together to revolutionize the way we are living in the future. To make it more accelerated and public friendly, we need to put research hours to bring the holy grail of deep artificial intelligence for practical real world engineering applications to ease human life. COMPANY DESCRIPTION/COMPANY OVERVIEW: SIGMIND is the very first Artificial Intelligence (A.I.) research group and engineering solution provider in Bangladesh. Our research group and developer team are working hard to invent ground-breaking solutions for complex problems to advance humanity. We are also collaboratively supported by NVIDIA Inception Program, Google Could for startups, Amazon  AWS Partner Network (APN) Global Startup Program to share different software and hardware resources and knowledge base.  The company is currently putting efforts to develop and market some innovative products with latest Computer Vision and Natural Lan

Abu Anas shuvom has been completed his graduation in Mechanical Engineering from Bangladesh University of Engineering and Technology (BUET) in 2016. During his final year, his research work has been published in SAE 2015 World Congress which is the largest Automotive Engineering conference being held in Detroit, Michigan since 1930. For his thesis work, he has also been recognized as a Highly Recommended Entrant in Global Undergraduate Awards 2015 which is the only pan discipline academic award program in the world. He is the first one from Bangladesh to be nominated for the award in Engineering category. He has also been served as an Engineering Panel Judge on that Award program in the following year. Since graduation, he has been walked in the path of entrepreneurship in association with Connecting Startups Bangladesh 2016 with the company he founded named SIGMIND. In his entrepreneurship career, he has been received numerous awards including Startup Bangladesh Award 2017, 2nd

We all know the immense pain of losing someone close to our heart to death. When the dead person also remains our Facebook friend, its hard to admit that they won’t replay back to them anymore. What if people could chat with the dead person on his/her “Remembering” page to better know about their values & interests? It might sound fancy, but technological revolutions enable us to make fictions into reality. The idea of “Virtual Self” isn’t very new, but the implementation would be significant. “Virtual Self” is a Deep Learning based technology which can replicate someones reply if trained on his/her chat history on social medias, specially Facebook. We have created a Character level language model with deep neural network which could be trained with anyones Facebook data, specifically chat history. After some parametric optimization, the language model works pretty well creating “reply message” to user queries by itself. The “Virtual Self” model could be furt

Being an enthusiastic and passionate researcher and developer in the field of Mechatronics and Robotics field, I had participated on different Robotics competitions including TECHFEST International Robotics Challenge arranged by IIT Mumbai in 2013 and placed 4th among 25 teams. On that competition , I was the team leader and developed a manual and an autonomous robot that worked synchronously to complete different objectives including solving a 5x5 grid to locate, pick and drop objects on the shortest path. The autonomous and manual robots had to work in synergy to maximize points while avoiding dead zones. I had written the whole embedded software code as the team leader which included small scale BFS algorithms for path planning. The code has been sourced at github (https://github.com/xhuvom/Grid_Solving_Robot) During my final year undergraduate courses, I had done a research as the lead author on improving automotive engine efficiency using machine learning metho