Vishal Babu

8.3+ years of experience in Data Science & Development. Expertise in machine learning (ML), deep learning (DL), natural language processing (NLP), Computer Vision, data science, artificial intelligence (AI), and software engineering. Worked extensively on gen_ai, LLM, CNN, RNN, LSTM, GAN, PySpark, sparkML, Tensorflow, Keras, PyTorch, NLTK, spacy, Stanford NLP, TPOT, PyCaret, RASANLU, Docker, Flask, AWS, object detection, object classification, and object tracking, etc. Seasoned Data Scientist experienced working with large datasets, breaking down information and applying interpretations to complex business concerns. Proficient in distribution, predictive and hypothetical modeling. Bringing [Number]+ years of related experience strengthening company operations.

• Object detection classification and tracking (ODCT), In this use case, we have created a real-time object detection classification and tracking system for BFL. This solution to identify treat by a camera and generate an alert. For object detection, we have done transfer learning over YOLO (You only look once) in darknet and trained a Deep-SORT (Simple Online and Real-time Tracking) in PyTorch on mars and Nvidia AI city dataset., YOLO, DeepSort, PyTorch, market 1501, darknet
• Anomaly detection in manufacturing, We developed a pipe-lined solution for detecting anomalies in manufacturing through computer vision and deep learning. The solution is based on Homomorphic filtering, and deep convolutional generative adversarial networks (DCGAN) and works well for less amount data as well., Homomorphic filtering, DCGAN, MvTecAD, GAN, synthetic data generation
• Ticket classification, In this project, our aim was to classify email based on the semantic information. And sent this email to the appropriate person. We were getting this email into multiple languages. So here we have written language detection, machine translation (Sequence to sequence RNN model), text cleansing, LDA and clustering., LSTM, Sequence to sequence, Spacy, NLTK, Text analysis, Stanford-NLP
• MEDS (India Automation), MEDS is a data collection product of Morningstar. Every organization has to submit their annual reports/quarter report to the Stock exchange (SE) in PDF format. Morningstar takes these reports then extracts Face of the statement (Income statement, Balance sheet, cash flow, and Notes) from these reports then fetch data points from the extracted face of the statements and compare these data with earlier reported data. We have Deployed ML models to automatically extract the face of the statementand the model for automatically extract data points from the table then matches these data point to earlier reported morning star data point using machine learning., Abbyy, SVM, random-forest TPOT, Autokeras, Keras, Flask, classification
• LTI Corp (Travel Claim prediction), Delivered classification model on employee travel claims based on the data using machine learning (SparkML and SK-Learn). Applied novelty detection, Isolation Forest and Box plot (Interquartile range) model to detect outliers using Machine learning (SK-Learn, NLTK, classification)., SparkML, SK-Learn, NLTK, classification
• Machine Learning on Mosaic Decisions, Mosaic Decision is a platform that serves business intelligence and analytics as a service. Integrated linear regression, decision tree regression models for Regression using machine learning (spark ML). Integrated Logistic regression, Decision tree classification, and random forest models for Classification using machine learning (spark ML). Integrated K-Means for clustering using machine learning (spark ML). Integrated correlation and dimension reduction for data understating and feature importance using machine learning (spark ML, sk-learn, and custom algorithms). For data preprocessing, integrated label encoder, one hot encoder, outlier treatment, null value treatment using machine learning (spark ML, sk-learn, and custom algorithms) For text mining integrated tokenizer (simple, regex and n-gram), word2vec, count vectorizer and tf-idf using machine learning (spark ML, sk-learn, and custom algorithms)., spark ML, sk-learn, custom algorithms
• Credit Information Report (CIR), In this project, we hit the Experian information bureau and fetch customer's financial information, based on this information we calculate credit score by a perimplemented scoring logic. It helps a particular entity (bank or financial institute) to make a relational decision on the customer. (Java, Python, ML)., Java, Python, ML
• Customer Centricity, This project was the POC for a client. In this project, we hit the bureau and get the important information like contact info, address info etc. Based on this info we trace the person on social media like twitter, e-commerce and try to predict the next possible activities of the person and sent then the best possible offers or suggestions. (Python, ML, Spark-ML, Stanford-NLP, Spark-Streaming), Python, ML, Spark-ML, Stanford-NLP, Spark-Streaming
• Version Control management system, At AutoRABIT Inc it was my first project. Here I have done research on version control management systems and integrated almost all major version control systems like GIT, SVN, TFS and Perforce in the AutoRABIT. Now user can do test connection, resister repository, delete a repository, create a branch, delete branch and import branch, etc. on the fly at any cloud by AutoRABIT (Java, git, svn, TFS, perforce, git hosted above aws, jQuery, HTML, CSS), Java, git, svn, TFS, perforce, git hosted above aws, jQuery, HTML, CSS
• EZ-Checkin, In this project, we have developed a module that connects Salesforce org to version control and synchronizes data between Salesforce org and version system. User are able to fetch the data according to specific commit, selective data or from head revision and commit to specific repo according to user need (Java, Version controls)., Java, Version controls
• User-Based-Auto Commit, User-Based-Auto-Commit automatically triggered at the time of CI jobs. It fetches the data from the sandbox or from one org to another org and maintains the history of users based on the users. (Java, Version controls), Java, Version controls

Noida (Vishal Babu)

Date of Birth: 11/10/89

Attended HiPC (High-Performance Computing) 2014 organized by IEEE and ACM at Cidade de Goa. Participated as a volunteer in ADCONS '13 organized by TEQIP II and IEEE Bangalore at NITK. Participated in two-day workshop on "Recent Developments in Software Testing at NITK. Participated in the workshop on Medical Image Computing at NITK.

I hereby declare that all the information furnished above is true to the best of my knowledge and belief.