Hi, I'm Neetesh Dadwariya
A
Self-driven, quick starter, passionate programmer with a curious mind who enjoys solving a complex and challenging real-world problems.
About
A Machine Learning Engineer with a passion for Deep Learning and Computer Vision.
I am a Computer Science Grad Student at The University of Texas at Dallas.
I enjoy problem-solving and designing solutions for unexplored problems. With strong Software Development industrial experience of 5 years in the customer experience industry, having a diverse skillset of engineering and big data pipelines, I am specializing in Data Science.
Detail-oriented Software Developer with professional experience in Backend Development, Machine Learning, Computer Vision. I am adept in Python, PyTorch, Machine Learning, Computer Vision.
Looking for an opportunity to work in a challenging position combining my skills in Machine Learning, which provides professional development, interesting experiences and personal growth.
- Software Development: Java, SpringBoot, Kafka, AWS, Elasticsearch, React, RESTFul APIs, Microservices, MySQL, MongoDB, Architecture Design
- Machine Learning: Python, Pandas, PySpark, TensorFlow, Colab, Scikit-learn, Databricks, Computer Vision, OpenCV
- AWS: S3, Lamda, SNS, OpenSearch, EC2, ECS
- Tools & Technologies: Maven, OpenTSDB, Apache Druid, Hive, Kibana, Grafana, Git, Docker, Linux, Jenkins, JIRA
Experience
Machine Learning Engineer
- Architected and implemented Machine Learning Pipeline using XGBoost for the detection of malicious commands running in Windows systems.
- Incorporated TF-IDF, Sent2Vec embeddings and performed Feature Engineering on PySpark. Processed Gigabytes of streaming data to identify anomalies in executed commands. Verified the binary classification metrices with 95% Confidence Interval.
- Utilized PyTorch and LSTM to create Deep Neural Network (DNN) model that successfully detected anomalies in streaming event data. Reduced false anomalies by 10%. Performed Cyclical Encoding on the day-of-week and hour-of-day features.
- Established Model Evaluation Pipeline for the newly built network anomaly detection models.
- Conducted SHAP analysis in SageMaker on deployed DNN model, with the goal of explaining its outcomes and providing stakeholders with insights into feature contribution.
- Tools: Deep Learning, AWS SageMaker, PySpark, PyTorch, Spark MLLib, Predictive Modeling, Forecasting, Statistical Analysis, Hypothesis Testing, Distributed Processing, Data Analysis, Databricks, AWS, Scikit-learn, Pandas, Python
Jan-23 - Aug-23 | USA
Data Scientist
- Leveraged statistical data science techniques including Hypothesis Testing and Confidence Intervals in Pandas-on-PySpark to perform Network Anomaly Detection on user login and logoff events from approximately 40GB of Windows Event Sensor Data, successfully flagging anomalies.
- Conducted Tenant-wise anomaly detection using the Time Series Prophet Model.
- Created user specific feature with data preprocessing based on the time series behavior for users, considering their login and logoff timings.
- Performed PCA on all features to extract relevant transformed features using dimensionality reduction.
- Applied Winsorization was employed to reduce the effect of outliers by scaling the data.
- Identified service accounts using K-Means unsupervised clustering algorithm to improvise overall detections improving detections by 20%.
- Tools: Python, Pandas, PyTorch, Databricks, Prophet Model
May-22 - Aug-22 | USA
Senior Software Engineer - II
- Experienced in the area of Software Design & Development of applications using Java technologies (Core Java, J2EE, Multi-threading, Collections, Microservices, JDBC, Object Oriented Principles) automating 60% of the manual processes.
- Developed a state-of-the-art platform to collect customer NPS and CSAT feedback from various sources. Built from scratch solution leveraged Spring Boot, Kafka, and MongoDB to streamline the workflow. With a visually appealing UI powered by ReactJS, integrated it across all business units, enhancing organizational efficiency.
- Spearheaded development and maintenance of backends for all post-booking Microservices, such as booking cancellation, change in travel dates, and purchasing flight seats and baggage reducing manual overhead by 80%.
- Contributed in entire SDLC process of Software development by adopting Agile and Test Driven Development (TDD) methodologies as Individual Contributor.
- Successfully delivered in-house projects on AWS Cloud utilizing SpringBoot, ReactJS, Kafka, MySQL, MongoDB, Neo4J and Docker containers. These projects included various automated features, resulting in a 25% reduction in ticket resolution time.
- Migrated all on-premise microservices to AWS ECS containerized environment.
- Built NLP driven chatbot pipeline to classify customer query intent and respond according to business use case.
- Tools: Java, Python, Maven, Springboot, DropWizard, AWS, MySQl, Kafka, Docker, ElasticSearch
June 2016 - July 2021 | Gurugram, India
Projects
Shoe brand image classification on Nike-Adidas dataset
Accomplishments
- Tools: CNN, Deep Learning, PySpark, OpenCV, Seaborn
- Constructed shoe brand image classification model on Nike-Adidas dataset using MapReduce based Convolution Neural Network.
- Performed training and inference on Spark Distributed Computing Environment.
- Incorporated Data Augmentation to enhance training data, rescaling. Experimented and logged various Hyperparameters, obtained model accuracy of 80%.
- Additionally, partially retrained model using transfer learning on ‘InceptionV3’ model to achieve model accuracy as 94%.
Transfer learning based flower classification