Machine Learning Engineer /quantitative Analyst Resume
Nyc, NY
SUMMARY:
- 15+ years of proven record of accomplishment in enterprise Software Product Development, Customization, and Implementation using C/C++, Java, SQL, Linux, Windows, Oracle, SQL Server, and JavaScript frameworks.
- 5+ years of experience in data analytics, machine learning, statistical data modelling using SQL, Python (tensorflow, keras, NLTK, panda, numpy), R Studio, MS Excel, and WEKA on desktop VMs and in cloud environment (AWS).
- Built server - side components, and data-based solutions as a part of systems integration and complex enterprise software.
- Led prescriptive analytics practice for projects on decision optimization using MS Excel and Analytics add-on toolkits.
- Balanced approach to solving business problem with attention to detail without losing big picture.
- Well versed with and hands on experience creation of data lake using
- Received recommendation and appreciation from a vice president of Moody’s while working on financial data analysis project.
- Thorough techno-functional leader with truly diverse experience developing and implementing enterprise application solutions for tier I financial institutions and engineering firms.
- Passionate about technology, statistics, applied mathematics and its applications to solve data-based business problem.
- Ability to influence and drive teams without authority and drive cross functional and cross departmental teams in the interest of overall quality of end-product.
- Extensively used what-if analytic tools like Solver and Treeplan in solving linear and non-linear optimization problems.
- Great appetite and interest in concept to reality projects in agile mode.
- Implemented decision tree algorithms in R and Python for binary and multi-class classification.
- Hands-on and working knowledge of Azure, GCP and Open stack cloud environments, set up and usage for machine learning algorithms implementation and applications deployments.
- Used ensembles to enhance accuracy of classifiers
- Proven ability to decide best evaluation metric and deliberate on trade-offs between choices for selection of right classifier.
- Proven ability to derive actionable insights from business operational data and make recommendations to principle statisticians and business leaders. Used mainly R-Shiny for visualization. Well versed with other visualization tools, MS PowerBI, and Tableau.
- Experience in data ingestion and storage of structured and semi-structured log data using flume and sqoop, for banking clients.
- Used TensorFlow and Keras libraries in Python for implementing CNN for image classification to implement proof of concept study on a security application where based on labels/names of people in trained data, access to be given to registered users.
- Clients/Companies worked for - HITACHI, Confidential, ORACLE, GE Financial Assurance, CITIBANK, Confidential, Confidential, Confidential, Confidential ’ Investor Services.
TECHNICAL SKILLS
Software: C/C++, SQL, R, Python (panda, keras, tensorflow, numpy, nltk), Core Java, MATLAB, R studio, Apache Spark, CSS, HTML5, Bootstrap, Jquery, SVN, Maven, Jenkins, angular, Sqoop, Flume, Kafka, Hive, ETL-Informatica
DBMS/NoSQL: SQL Server, Oracle, PostgreSQL, MongoDB
Machine Learning: LDA, PCA, Text Analytics, Regression, Classification, Hypothesis Testing, Cluster Analysis, statistical modeling, CNN, ANN
Operations Research: Decision optimization, Linear, Integer and non-linear
Data Visualization: R Studio, matplotlib, Microsoft PowerBI, Tableau
PROFESSIONAL EXPERIENCE:
Machine Learning Engineer /Quantitative Analyst
Confidential, NYC, NY
Responsibilities:
- Worked with vice president of Moody’s. Studied Coating Industry and top five players in industry.
- Presented comprehensive analytical report and recommended bond ratings to investors, for a coating industry.
- What-if analytics on financial investment data to recommend various deal options and impact analysis on bond credit rating. (Solver, MS Excel)
- Designed interactive visualizations to track financial performance of major paint and coating companies. (Tableau)
- Created dashboards such as social media brand presence, financial activities, competitor performance. (Tableau)
- Data collection, processing, and reporting framework. (web scraping, Beautiful Soup/Python)
- Ingestion, normalisation, and analysis of marketing intelligence data relevant to selected paint and coating companies.
Machine Learning Engineer
Confidential, New Brunswick, NJ
Responsibilities:
- Uncovered actionable insights from unstructured quality data from Surveillance Risk Management System.
- Worked with Principle statistician and made recommendations that drove business decisions to improve quality of operations in the unit.
- Visualization, evaluation of clusters with similar device complaints.
- Statistical modelling for prediction of resolution category, prediction of root cause, prediction for average resolution time for the quality related issues.
- Pre-processed unstructured complaints data to convert it into document term matrix.
- Using rtexttools library, applied classifiers to predict complaint resolution category and root cause.
- Developed forward feature selection algorithm in MATLAB
- Implemented topic modelling concept on data records to predict resolution category.
- Implemented word2vec using python library functions, to derive word vectors associated with specific resolution category and root cause.
- Exploratory data analysis of device complaints operational data with 50k records. (Tableau)
- Exploratory data analysis of patient data of size 10+ GB. Data wrangling using PostgreSQL, visualization using Tableau and predictive modelling using Logistic regression.
- H2o.ai library, R, Python -NLTK, Word2Vec, NLP, topic modelling
Senior Software Consultant/TPM
Confidential
- Data warehouse design using hive for devices data for batch processing.
- Defined staged pipeline for batch processing, from ingestion to visualization for exploratory data analysis.
- Designed and documented process for streamed data processing using spark.
- Performed digital market mix modelling, to provide an optimal budget allocation by formulating revenue as a function of marketing spend and building S response curves that identify minimum and saturation spends for an advertising medium
- Built marketing spend optimizer using nonlinear optimization technique and leveraged market mix models to recommend marketing spend across geographies and channels
- For a bank in Europe, Mentored team of three developers and done reviews for processing semi-structured log files data to:
- Develop hive UDF’s and UDTF’s to transform the data suitable for analysis.
- Customize InputFileSplit and FileInputFormat classes to process different file formats on HDFS.
- Partitioning and Bucketing hive tables.
- Automate the Workflow process using Oozie.
- Migrate the existing database to hive using sqoop.
- For a Telecom client Built a recommendation system using Market Basket Analysis on the subscription data:
- Which helped them to make customer centric recommendations.
- Extracted the streaming data using Apache Flume
- Built the final reports using tableau Dashboards.
- Proof of concept study for an energy utility company. Data store preparation and maintenance. Worked on ingesting, munging, joining business data to predict customer churn probability. Hadoop infrastructure capacity planning.
- Led efforts for the development, customization, and implementation for consumer banking suite of Finacle for tier-I banks across globe.
Confidential
Senior Software Consultant/TPM- Led development efforts to incrementally add functional features to internet banking stack of Finacle.
- Implemented enhancements in Responsive UI development to make multiple device and browser compatibility.
- Integrated IBM Trusteer, cloud-based fraud detection analytics solution with ANZ internet banking.
- Made internet banking simple, with rich user experience by introducing simple way to enter payment details including BPAY, Pay Anyone, Transfers, and international services.
- Enhanced business functionality of transaction history by introducing new search option based on amount range of records.
- Java, SQL, angularJS, JQUERY, Oracle, Maven, Jenkins, front end web design tools for responsive UI(CSS, bootstrap, HTML)
Confidential
Senior Software Consultant/TPM- Led core banking business transformation for a program of Department of Post ( Confidential India).
- Oversaw, governed, and mentored customization design, development around base product, requirement and solution scope management, configuration, and deployment management.
- Designed and developed user hooks / APIs for consumption by internet banking team.
- This development has laid strong foundation for further success in pilot and staged role out of the transformation.
- Legacy scripting, JavaScript, SQL, Oracle, UNIX, SVN
Confidential
Senior Software Consultant/TPM- Business transformation involved heavy integration with peripheral systems.
- Requirements gathering, data modelling and data flow mapping.
- Batch integration for business transformation of deposits module - Retail banking
- Design and development of database for batch interfaces in data warehousing module. Architecture and code reviews.
- Extract transaction history data from legacy system.
- Design data pipeline from mainframe to Unix environment. History data migration and processing strategy and execution
- Deployment, and configuration management process initiation, set up and evolution.
- Data modelling and data flow mapping. Batch integration for business transformation of deposits module - Retail banking. Architecture and code reviews.
- Design data pipeline from mainframe to Unix environment. History data migration and processing strategy and execution
- Cloud Deployment and configuration management of device management and monitoring products.
- ETL and data migration.
Lead Programmer/Analyst
Confidential
- Design, development, implementation and go live support of fund transfer system for Citibank regions (Asia pacific AU/NZ and Canada).
- Wrote server-side interface components as part of C++ dynamic link libraries in UNIX environment.
- Conducted requirement gap analysis, design, development of upstream and downstream interfaces of partner systems with Global Pay Plus (GPP).
- Straight through processing of SWIFT messages, rules set analysis, static data set up and configuration of GPP.
- SIT and UAT support, production issue analysis and resolution. Analysis and solution of messages in repair.
- Extensively worked on designing and writing interfaces that were called by core GPP objects SWI, SWX, RTR, HRQ and CPA.
- Functionalities of interfaces included
- Batch interfaces for uploads from GL systems and other partners.
- Memo posting
- Creation for data feeds
- Querying external systems
- Advice generation
- Payment messages upload through files
- Global interdict check
- Querying to external systems
- Prepared functional and technical documentation to mentor and train testing and development teams.
- Technology - Visual Studio C++, SQL, PL/SQL, UNIX, Tuxedo
C/C++ Developer
Confidential
- Coded for adding 64 - bit compatibility for USB auto-detect installation feature on windows, UNIX, and LINUX for processors AMD64, EM64T and Intel-ITANIUM in SNMP, serial, and USB mode.
- Coding and fixing to Port of LANSAFE (Power Management Software) on 64 BIT operating systems (Windows, LINUX, and UNIX).
- Used byte-swapping libraries in C/C++ to solve issues arising out of network byte ordering on LINUX and UNIX servers.
- Received valuable contribution award from CEO of Confidential for successfully starting development efforts from Pune development center for set of products.
- Technology - C/UNIX, MFC, Socket Programming
Java Developer
Confidential, Virginia
- Developed web-based data warehouse tool in JSP/Servlets, HTML, JDBC that effectively acted as a dashboard for executive leaders to track weekly and daily progress.
- Incremental feature additions in the dashboard and weekly deployments in iplanet application server and weblogic web server in agile development environment.
- Worked with DBA to extract required data from insurance subsystems and plan refresh of data at regular intervals.
- Led team of two developers to design and develop multithreaded GUI based FTP tool in Java, using Swing library for user interface. It acted as a routing switch for continuous file transfer to partner systems.
- Improved performance of the Java FTP tool by adding capability to auto restart the thread in case of the exception caught in the polling and FTP process.
- Developed modules for database connectivity to Oracle policy database.
- Designed and developed classes and methods to in corporate business logic, new policy number generation and assignment to policies.
- Wrote and optimized SQL queries to design materialized views as part of business logic to perform efficient policy search and retrieval operations on oracle database.
- Orchestrated all stages of data pipeline considering refresh times, extract, transformation, and load operations for subsequent data processing.
- Reduced cost of operations by decommissioning third party tool that was used previously to perform policy search and view operations.
- Led development efforts to design and develop an image processing tool that was used in creation of index files and merged image files of individual policy documents.
- Used Lead tool libraries in Visual Studio to merge and process images.
Environment: Core Java, SQL, SQL Server, JSP/Servlets, Swing, AWT, FTP classes, JDBC, Windows 2000, WinNT
C++/CORBA VisiBroker Tester
Confidential
- Study VisiBroker and CORBA specifications.
- Identifying test cases, Creating and Reviewing test cases in Program Checklists (PCLs) for the product.
- Writing and Reviewing C++, Java-CORBA stubs skeletons for testing functionality on Solaris, Windows, and HP-UNIX Platforms.
- Test the product on Solaris, Windows, and HP Platforms.
- Prepare Software Problem Report.
- Environment: C++ 4.0 and 4.1, JDK 1.3.0, JDK1.2.2, Java, VisiBroker 5.0, Solaris 7.0, Solaris 2.5.1, Solaris 2.6, HP UX 11.00, Windows NT, AIX 4.3.3
C++ Developer
Confidential
- Wrote C and C++ library APIs in Windows environment, for consumption in the implementation of new AutoCAD features in HICAD.
- Using existing C/C++ APIs in WINDOWS environment, developed algorithm to allow specification of drawing origin by user and to allow new features to be configured in drawing format.
- Developed an algorithm in C to allow user to choose color code conversion mapping.
- Designed and developed user interface in Microsoft Foundation Class framework.
- Wrote high level and detailed design specifications for information module of DXF Converter.
- Studied current MFC Document View Framework and developed an algorithm in VC++ to modify CLSUM function to exclude certain checklist items of checklist editor CLMAN. CLMAN was a proprietary MS EXCEL like product.
- Change control of entire source code using VSS.
- Followed SDLC cycle to implement these projects.
Environment: C++, VC++, MFC, Hi-CAD draft, Windows 95, 98, 2000, WinNT
VC++/COM Trainee Developer
Confidential
- Hands on training on VC++, Java, Oracle, SQL and ATL COM (component object model).
- Functional and technical documentation of as-is TBA systems (Total Banking Automation) developed in legacy.
- Documentation of to-be systems to integrate bank branches with central control application.
- Message format preparation, and documentation of message flow.
- Data modelling for local and central databases.
- Developed dynamic link library objects (APIs) in ATL COM framework.
- Designed, developed, and tested middleware components having business logic (transaction initiation, tracking, GUID creation, user validation, audit log) and database connectivity.
- Unit, system and integration test case preparation and execution.
Environment: C++, VC++ 6.00, COM (ATL), Windows NT 4.0