Manisha Biswas is an author, blogger, data scientist, and a community person based in Buffalo, New York. She is currently pursuing a Masters in Computer Science and Engineering from
The State University of New York, Buffalo majoring in Machine Learning. She exhibits good leadership skills and has led independent community programs to empower women with technology by hosting regular technical workshops, collaborating in groups to work on different technologies. She is always very passionate and eager to share her knowledge and ideas with others. This passion is what inspired her to author three books- Reinforcement Learning With Open AI, TensorFlow, and Keras Using Python, Neural Networks in Unity, Beginning AI Bot Frameworks Getting Started with Bot Development. She believes that with passion, hard work, gratitude, and positivity magic can happen. She loves traveling. Having been to different places and connecting with people from different cultures she realized that we all are more alike than we are different.
-
Experience
I have a rich experience of 2.5 years working as a full-stack developer and as a data scientist developing products and implementing strategic initiatives. Skilled in various technologies including Web Development, IoT, Object-Oriented Design, Machine Learning.
1. Optiks Innovations Pvt. Ltd. | Data Scientist (Dec 2017 - Oct 2018)
• Worked in a pharmaceutical solutions project designing and developing an application bot to answer specific business logic questions for US-based clients using C#, LUIS, Microsoft Bot Framework, Visual Studio, Microsoft Azure.
• Integrated Tableau Rest APIs with the bot to show tableau reports as images based on user queries.
• Responsible for implementing an end to the end life cycle of a software project starting from design document creation to post-implementation support.
2. Insync Tech- Fin Solutions Ltd. | Software Developer (July 2016 - Nov 2017)
• Developing a studio (AppseConnect Workflow) to implement diagrammatic workflows and automate the data and control flow of integration among various applications such as eCommerce, CRM, and ERP using C#, Visual Studio, Go.js, Knockout.js, SQL Server, Microsoft Azure.
• Designed a rule engine to provide decision making capability to Workflow by which user can create a rule specific to their business needs and trigger an action like email, raise a support ticket
• Created logic elements to split, merge, or aggregate data directly and halt, resume, or schedule processes within the workflow.
• Workflow helps the user to graphically create, view and change the data flow for a process by quickly dragging and dropping the logic elements and automate their data integration process
• Developed responsive User Interface for AppseConnect Cloud Portal using HTML5, Bootstrap, and CSS3.
• Worked in Agile Methodology to communicate and demonstrate the product features to the product owners.
• Co-ordinated with the QA team during UAT cycles to prepare test reports.
3. Apress Publications | Author (Dec 2017 - Present)
Author of books -
• Beginning AI Bot Frameworks: Getting Started with Bot Development
• Neural Networks in Unity:C# Programming for Windows 10
• Reinforcement Learning with Open AI, TensorFlow, and Keras Using Python
-
Projects
1. Audio Situational Awareness Device
Technologies Used: Python, Java, ReSpeaker Core v2.0, Android Studio, GCP
The ASA device is intended to help deaf people listen to anybody and engage openly in group discussions as it involves working with a hardware platform i.e the ReSpeaker Core v2.0 which captures speech and expands it to work with Google Translation API to provides the speech to text translation functionality. The speech is then displayed as a text message on an Android-based application. Also developed an ML model to uniquely identify a speaker based on frequency spectrogram, word utterances, etc. by gathering samples from a set of speeches from the speaker audio.
2. Drug Discovery and Repurposing
Technologies Used: Python, IBM RXN, DeepChem Library
Developed a ML model to identify new drugs and test its toxicity. Performed a comparative study on the TOX21 dataset implementing different algorithms, such as. Graph Convolutional Network, Generative Adversarial Network, and received a satisfactory result of 99% for Graph Convolutional Network.
3. Electing World Bank President
Technologies Used: Blockchain, Remix, Solidity, Truffle, Ganache,Web3.js
Developed a decentralized end to end voting application to elect world bank president with Upvote and Downvote feature.
4. Movie Genres Prediction Based on Plot Summaries
Technologies Used: Python, Pyspark, CMU Movie Summary Corpus dataset
Worked on a multi-label classification problem and developed a model for predicting movie genres based on plot summaries provided in CMU Movie Summary Corpus dataset using PySpark. The model achieved a satisfactory result to predict all the genres associated with the movie.
5. Big Data Processing with Hadoop
Technologies Used: Python, Hadoop, MapReduce
Implemented basic text processing tasks from scratch on the
Hadoop framework such as implementing a MapReduce algorithm to calculate wordcount, tri-grams around the keywords, produce an inverted index for the customized dataset, join two datasets using a primary key.
6. Binge-Watch: Alexa PC skill
Technologies Used: Python, C#, Amazon Alexa Skill for PC, AWS, Visual Studio
The 'Binge-watch' is a PC-based Alexa skill that recommends movies based on a person's previously searched or viewed movies.
7. Path Learner
Technologies Used: C#, Python, Unity Game Engine, Unity ML agents toolkit,TensorFlow
Trained a simulated environment using Reinforcement Learning approach in Unity to work on its own in the most efficient manner.
8. Samurai AR
Technologies Used: C#, Unity Game Engine, Vuforia, Android Studio
Developed an app that uses a smartphone camera as an AR camera to display a 3D model ov