SUMMARY:

• I can start to work any employer in the US anytime within a week.
• Expert in databases to extract, transform and load data to analyse and get interpretable/meaningful result
• Expert in building predictive models for both regression and classification problems.
• Expert in Natural Language Processing.
• Expert in Unsupervised modeling (KMeans clustering); find patterns in data, but without a specific prediction task in mind. Transform features for better training, crosstab to check performance of model.
• Operational research methodologies and implementation over the projects. Such as having a comprehensive propensity score estimation model which is a better alternative to logistic regression based on multiple operational research methodologies on one of my past project.
• Database Developing, design and implement schema, data migration, Hadoop integration, Hive, sqoop on claudera to migrate data between RDBMS and Hadoop.
• Data mining, data cleaning, data analysis, statistical machine learning predictive modeling, NLP modeling.
• Ensemble Modeling (voting classifier, bootstrap/bagging)
• Deep learning with keras.
• PySpark big data fundamentals, PySpark Feature Engineering, Pyspark machine learning pipelines, Pyspark model tuning and selection.
• Apache Airflow ETL on google cloud platform, workflow as a DAG(directed acyclic graphs) to schedule retrieve data, run big queries and periodically train models and report.
• Connecting tableau to different data sources, and apply real time reporting and visualization.
• Google big query, ad hoc reporting, tableau, plotting.
• Hypothesis testing, A/B testing. Data transformation, Data normalization
• PCA (Principal Component Analysis) in Python and R.
• Apply regression (tensorflow,keras,sklearn), logistic regression (tensorflow, keras, sklearn), ElasticNet(sklearn), XBoost(sklearn), random forest classifier (sklearn), n - gram tokenizing NLP (sklearn), Support Vector Machine - SVM (sklearn, tensorflow, keras), k-NN classifier (tensorflow, keras, sklearn), Tree-Based Models (sklearn, tensorflow, keras), XGBoost ( Extreme gradient boost in sklearn), SGDC (Stochastic gradient descent classifier in sklearn). KMeans (sklearn, tensorflow, keras), Naïve Beyes for text analysis
• Apply sklearn pipeline for automated machine learning with cross-validation in GridSearchCV and hyperparameters tuning.
• Supervised learning workflows (deploying models in production), feature engineering across multiple data sources, productize the model and ensure it continues to perform well thereafter by iteratively improving it.
• Diagnose dataset shift and mitigate the effect that a changing environment can have on accuracy.
• Modelling data without any, or with very few labels (kind of unsupervised learning which takes us into anomaly detection).
• PCA (principal component analysis) for fundamental dimension reduction (intrinsic dimension) to increase efficiency in storage and to decrease computation.
• GIT a modern version control tool that is very popular with data scientists and software developers alike.
• Querying HBase with Apache Phoenix.
• The universal glue of programming Unix Shell Script > batch processing.
• Scalable machine learning stack: APIs, containers, running.
• Canary deploy: rollout a new model with a small subset of users and then monitor its performance for a while.
• Blue/green deployment make sure the model is compatible with fully production
• Model as a service AWS, Azure and Google Cloud Services
• Horizontal scaling
• Microservices:
• Deploying a machine learning model with an API,
• Run the model and API in a container,
• Manage the containers using an orchestration service,
• RESTful API: call over http and return JSON

PROFESSIONAL EXPERIENCE:

Confidential, Illinois

Android iOS Developer

Responsibilities:

• Integrated location-based services using Google Maps API from the GPS to display locations of the nearest ports
• Collaborated with web services team to implement the backend support.
• Worked on linking the FedEx Web Services Android App using Faster Jackson.
• ImplementedAndroid Bluetooth and Location Connectivity components.
• The usedMockito framework which helped to write tests with a clean & simple API.
• Implemented Google's Material Design for better look and feel of user experience.
• Implemented ListView, to show the different shipments and maps as a form of a list
• Implemented SQLite databases to monitor shipments.
• Push Notifications from an MQTT message broker server.
• Worked briefly on the complicated AOSP build system to determine the viability.
• Worked extensively on Eclipse Kepler.
• Debugged/Integrated/Maintained a Trusted Execution Environment (TEE) for High-Security OMAP devices enabled with MShield technology for newest Android HLOS and kernel.
• Debugged/Identified kernel and ROM bugs, related to MPUSS and power management
• The app using the location-based services demanded integration of Wi-Fi, GPS, Camera and Bluetooth Integration.
• Completed a relatively simple work-in-progress written with XamariniOSnative tools (not XamarinForms).
• Worked extensively with Objective C and Cocoa frameworks.
• Implemented the NDK for the smooth functioning of the location-based services on Android.
• Implemented push framework on Google's push notification service.
• Also Implemented MQTT Client that is used to start a background thread responsible for sending a ping message to the server to keep the connection alive.
• Designed & developed front-endGUI and the hardware feedback control from the user.
• Expertise in Java Native Interface (JNI), used to access interface written in C++ and C.
• Used various caching APIs which enabled offline storage and helped to read data flawlessly even if the network is not available.
• Worked with UI Action Sheets, UI Tab Bar Controller, UI Table Views, Custom Cells, UI Scroll Views, Navigation Controllers, delegates and protocols to better the menu tab and smoother scrolling between tabs.
• Worked with UIActionSheets, UITableView, Custom Cells, UIScrollViews, Navigation Controllers, delegatesand protocols.
• Worked with Objective C, Swift and Cocoa frameworks.
• Implemented customized web view component to load an SWF file and HTML data.
• Worked briefly withOpenGL ES.
• Created test cases and test data for testing the application using JUnit Extensions
• Performed Unit testing on the various modules/functions to ensure that the output is as expected
• Implemented Recycler view in our existing Android application and working on various other features including Material Design.
• Leveraged the Android Bluetooth Stack used for pairing mobile handheld devices with a Bluetooth enabled printer via Serial and BSP communication ports.
• Tested the application on Robotium, Appium and Roboelectric, JUnit to ensure quality control of the app.
• Provided support for the desktop version of the app where Backbone.js was implemented on JavaScript framework with a RESTful JSON interface, based on the model-view-presenter (MVP) application design paradigm.
• Experience with Internationalization APIs like Formatters, Collation, and Message Format.
• Experience with localization to localize the OS and all the applications.
• Used JIRA extensively to keep a track of the many issues.
• Experience with Linux driver concepts.
• Also worked on other apps fleet in debugging complex issues that other traditional applications are unable to observe, worked on persistent logging apps to search for various hardware and communication problems from service providers, device hardware quirks and other client-server anomalies.
• Used cloud-based service database for Android - Firebase.
• Used Android SDK and NDK to well suited for the different kinds of Android Devices.
• Responsible for rendering the 2D graphics to the user using OpenGL.
• Also worked on the iOS version of the app., co-coordinating on the app code.

Environment : Android SDK, Eclipse IDE, Google Maps, GPS, Wi-Fi, AML and SQL Databases, JIRA, DOORS, JUnit, GPS Services, Appium, Robotium, Roboelectric, Bluetooth, Backbone.js, TDD, Swift, XCode, Android Kernel, Mockito, Bluetooth, OpenGL, Internationalization and Localization, HTML 5.

Data Scientist

Confidential, Milpitas, CA

Responsibilities:

• Gained experience working in an agile environment,
• Leveraging open source software and practicing agile software development
• Have small projects through all phases of software development from design to production deployment.
• Used the latest technology of stack, end learned best practices of software development in real world setting.
• Worked with alongside bright engineers.
• Understand the business objective is very crucial in order to work with right data
• Locate the necessary data sources in/out of the business
• Retrieve data thru relational (SQL) or nonrelational (noSQL) databases to Hadoop environment
• If data is on the web apply web scraping
• With ETL tools, such as mysql workbench, sqlserver SSIS, if data is very large and speed is very concern then move data to hadoop environment with ETL tool like sqoop, if schedule necessary apache airflow ETL tools is very documented tool.
• PySpark over the Hadoop or Hive queries over the Hadoop environment to drastically speed up the expletory data analysis process which takes the majority of the time in total project.
• Check for data clarity, then clean up, tidy up, apply some statistical fundamental tests (2 sample T test, one-way anova, 2-proportion test, chi-square, A/B testing);
• Check for normality, kurtosis, skewness, variance
• If necessary, normalize data to avoid units and scale issues
• Then apply some hypothesis test relevant to business objective
• Research methodology to find best model or models for the business objective
• Principle component analysis to reduce data dimension
• Establish the machine learning pipeline
• Train and hyperparameter tuning
• Check for accuracy over the test data
• Finalize the model and expose the model through an API
• Putting the model and API (RESTfull API) in a container
• Establish docker environment (install necessary packages for model to run)
• Blue/green deploy, or canary deploy kind of last exit before the bridge
• Kubernetes to dockerize and implement container technology to orchestrate model service in the cloud so manageable increased/decreased web traffic.
• Machine learning as a service
• Document the model’s API

Environment: Python 3.x, Linux, Tableau Desktop, SQL Server 2012, Microsoft Excel, Scikit-Learn, Keras, Apache Airflow, Apache Sqoop, ETL, SQL Server SSIS, mySQL workbench, Hadoop, Hive, Spark (PySpark), Kubernetes, Docker, RESTFul API

Computer Lab Assitant

Confidential

• Taught SQL, Python, noSQL, C++ to students.
• Database Developer, IT department, Ez inc Group. Kayseri, Turkey
• Designed and implemented new database schema for new management system, OLTP.
• Migrated existing data from sql server 2008 to mysql with migrate wizard in mysql workbench.
• Created databases, new tables and transferred existing data into new tables. Installed LAMP, and build ETL for data warehousing to have top down data marts.
• With new management system that covers manufacturing, production rate gained 25%, production loss decreased to 8% from 10% with the same number of employees, paper usage decreased to 1/4.

Endivronment: Mysql, Linux, SQL Server 2008, mysql workbench