We provide IT Staff Augmentation Services!

Machine Learning Consultant Resume

4.00/5 (Submit Your Rating)

Charlotte, NC

PROFESSIONAL SUMMARY:

  • Creating text and image based machine learning models using Advanced Python Libraries like Keras and Tensor flow
  • Experienced in writing production level python code
  • Familiar with GIT, AWS, Docker and Kubernetes technologies
  • Project management and leadership skills.
  • Wrote SAS codes to build predictive models and B2B marketing analysis on millions of rows of data in past job.
  • Expert knowledge in Data mining and Predictive modeling using CRISP - DM methodology.
  • Good in statistical analysis - Hypothesis tests, ANOVA, Experimental Design, Correlation and Regression.
  • Proficient in Decision Trees, Linear and Logistic Regressions, Neural Networks, Segmentation, Clustering concepts.

TECHNICAL SKILLS:

Analytical Tools: Jupyter Notebook Python 3.6, Pycharm, Databricks Spark 3.2, SAS 9.2, SAS Enterprise Guide 5.1, SAS Enterprise Miner 12.1, SPSS Modeler, Tableau

Databases: Teradata SQL, Oracle in UNIX platform, MS-SQL

Programming: Python, SAS, PL/SQL

Workflow tools: MS-Project, MS-Excel, MS-Visio, MS-PowerPoint and MS-Word

Project Management: Critical Path Method, Gantt Chart, Requirement Analysis, Work Breakdown Structure, Change Management

PROFESSIONAL EXPERIENCE:

Confidential, Charlotte, NC

Machine Learning Consultant

Responsibilities:

  • Working in Voice of Customer AI
  • Doing NLP on speech transcripts from Nexidia in Python
  • Built Random Forest Classification to identify if a complaint call is high risk or not.

Confidential, Park City, UT

Machine Learning Consultant

Responsibilities:

  • Used NLP techniques on social media data to perform sentiment analysis
  • Familiar with ngram analysis and other NLP techniques like stemming and lemmatization for effective text processing
  • Developed model which dynamically calculates confidence scores for events detected by the system using LightGBM
  • Built text models using Keras - Multi-class text classification to predict probability for different categorical levels
  • Built CNN+LSTM text model to get probability on social texts of how likely event is happening.
  • Built Random Forest Classification for Text Filter Model - Model that identifies texts that has be filtered out based on text features.
  • Used Word Mover’s Distance to compute text similarities.
  • Developed Support Vector Machine based Day/Night Image classifier
  • Used Uber's Kepler tool for advanced visualization of Waze traffic data

Confidential, Midvale UT

Data Scientist

Responsibilities:

  • Used Random Forest Classification in sklearn to build look alike models - Data exploration, Outlier Analysis, Imputing missing values, Correlation Tests, Multicollinearity checks, Model building and hyper-parameter tuning. dentifies customers having similar profiles to any of the current ClubO customers.
  • Identifies customers having similar profiles to any of the current Confidential cardholders.
  • Grouped similar profiled ClubO customers using clustering to create segments and used classification method to identify non-ClubO customers having similar profiles to the ClubO segments. Used PCA, K-means clustering and Random Forest Classification.
  • Built a regression model to predict an optimal date to send reminder email to buy again same product from Health and Beauty store.
  • Used Lasso Regression in spark Databricks.
  • Used Logistic Regression technique and wrote SAS codes throughout - Data exploration, creating derived variables, Outlier Analysis, Data Transformations, Collinearity tests, Variable reduction, Model building and evaluation
  • Predicted probabilities for customers’ response to a social media campaign. Customers who have been targeted through social campaigns in the past were taken with both audience and responders.
  • Predicted probabilities that a customer would buy from a store where they have not already purchased. Logistic regression model is built for each store to predict probabilities at customer level.
  • Used ARIMA method in SAS to predict weekly Nectar forecast.
  • Supported ClubO by tracking various ClubO program performances and suggested on improvements.
  • Conducted AB testing and profiled customers with all channel related metrics.
  • Monitored channel growth in terms of ClubO signups and profit.

Confidential, Princeton, NJ

Database Marketing Statistical Analyst

Responsibilities:

  • Used Logistic Regression technique and wrote SAS codes throughout - Data exploration, creating derived variables, Outlier Analysis, Data Transformations, Collinearity tests, Variable reduction, Model building and evaluation
  • Developed a targeting model for the Confidential Small Business and Mid-Market Acquisition process that predicts the probability of a caller response.
  • Developed a targeting model for the Confidential Small Business and Mid-Market Acquisition process that identifies prospects that has similar profile to Confidential customers.
  • Produced response curves for the period of response based on number of days to make first/last call.
  • Created detailed analysis report based on changes between the present and past results.
  • Analyzed service utilization (voice, data, video, transport and other) for each industry.
  • Performed Penetration analysis by calculating penetration and capture rates.
  • Performed Lift Analysis with varied incremental response rate and conversion rate.

Confidential

Oracle Database Developer and Team Lead

Responsibilities:

  • Wrote PL/SQL queries and developed views, functions, triggers and stored procedures on UNIX platform using SQL*Plus.
  • Handled 700 databases with 10TB of data in Oracle 9i, 10g and 11g and managed tablespaces, indexes and user profiles.
  • Took Backup using exp/imp and RMAN.
  • Managed a team of 25 system engineers by allocating tickets, scheduling work time, conducting Knowledge Transfer sessions and reported to Database Tower Lead.

We'd love your feedback!