PyModeltime

Lead Co-Author with  Matt Dancho  pymodeltime,  Python package for time series machine learning  

ChatOpenLLM

Developing ChatOpenLLM, a Python package built on the robust Langchain library that provides ChatOpenAI()-like functionality for various open-source models

Certifications in Big Data Analytics

1. Data Science for Business Part 1 

 Business Science University,  November 2022

2. Machine Learning and Big Data with Macroeconomic Applications, February 16 -17, 2019

Centre for Monetary and Financial Economics (CMFE), Carleton University 

Trainer: Dalibor Stevanovic 

3. High-Performance Time Series 

 Business Science University,  August 2022

4. IBM Data Analyst Professional Certificate

IBM, Coursera, May 2022

5. Analyzing Big Data with Microsoft R 

Microsoft, Edx, October  2017 

Testing and Evaluation of modeltime.h2o R package 

As a volunteer (Please check the Acknowledgements of the R blog on 

Introducing Modeltime H2O: Automatic Forecasting with H2O AutoML)  

 Data Science Skills through writing research papers

1. Expertise in recent advanced Large Language Models, including GPT-3.5, GPT-4, Vicuna, Falcon, Llama and other open source LLM. 

2. Proficient in using workflows like LangChain and Semantic Kernel to build LLM applications.

3. Applied advanced NLP models such as BERT, RoBERTa, SBERT, and RAKE algorithms and text mining approaches such as a grammar-based search for constructing an economic policy uncertainty (EPU) index (using Python)

4. Used deep learning models (LSTM, RNN) and machine learning models (XGBoost, LightGBM, CatBoost, GBM, Random Forest, Local linear Forest, GLM, and Prophet algorithm) for forecasting (using R and Python)

5. Automating the process for downloading the Google trends data (using R)

6. Stacked Ensemble

7. Timeseries machine learning models and econometrics models for forecasting

Shiny App

1. Forecast Canadian GDP Q-4 2022  using Prophet, NNET, and MARS

2. Loan Default Scorer  using XGBoost and TabNet  

3. Predict Customer Lifetime Value with Machine Learning  using XGBoost 

4. Combine PDF-Online PDF Combiner 

5. Customer survery  

Streamlit App

1. Resume-Job Description Similarity-S-BERT model 

2. Automation of newspapers count total and EPU Boolean for Canada 

3. Highlight Canadian economic policy uncertainty index (EPU) related words 

4. Automating Estimation EPU-NLP for Canada 

5. Semantic Search Engine using S-BERT model(It will take time to run)

Replication codes 

List of Labs done in Business Data Science

Business Science University,  March 2021 to date

1. Developed a framework for understanding the travel cost between vendors and distributors(any Geospatial problem), Geospatial Networks with sf, nngeo, & osrm (using R)

2. Customer Lifetime Value (CLV, RFM) with Machine Learning (using R and Python)

3. The non-linear models for price elasticity use hierarchical modeling for product groups and probabilistic modeling using Markov Chain Monte Carlo(MCMC) stan, Price Elasticity, Non-Linear Models (GAMs) & Hierarchical Models (using R)

4. Anomaly Detection with H2O Machine Learning (using R)

5. Time Series Anomaly Detection with anomalize on Energy data (using R)

6. Modeltime (ARIMA  & Prophet), Energy Forecasting Report Automation (using R )

7. Automated Marketing Mix Modeling (MMM), Facebook Robyn

8. Network Analysis For Customer Segmentation  (using R and Python)

9. Using Customer Credit Card History to Cluster with Network Analysis (using R and Python)

10. Customer Churn Survival Analysis with correlation funnel, parsnip, & H2O and Explaining Machine Learning for Customer Churn (using R)

11. Market Basket Analysis & Recommendation Systems with recommenderlab (using R )

12.  A/B Testing for Website Optimization with Infer & Google Optimize (using R )

13. Advanced Customer Segmentation  & Market Basket (using R and Python)

14. Hierarchical Forecasting and Feature Engineering for Customer Analytics (using R and Python)

15.    Airline COVID Forecasting,  Modeltime  &  Modeltime GluonTS model (using R and Python)

16. Energy Demand Forecasting using Autoregression Modeltime Recursive (using R )

17. Marketing Multi-Channel Attribution with Channel Attribution (using R )

18. Risk Analysis & Simulation with R | Shiny Monte Carlo Simulation App

Kaggle Profile

Technical Skills

Data Science and Analytics: Skills and Tools

Programming Languages and Frameworks: 

Python: EconML, Sentence-Transformers, Transformers, Statsmodels, NumPy, Pandas,  Scikit-learn, Keras, Tensorflow,  etc.,  R  : Timetk, Modeltime ecosystem, Tidyverse, Tlverse Software Ecosystem,  etc. Pyspark (Python API for Apache Spark), H2O Driverless AI, SQL, Tableau, Google Data Studio, Power BI, IBM Cognos, AWS Cloud, Saturn Cloud, Model Deployment for Production with AWS  and the basics of  Web Scrapping.

Machine Learning Skills:

Supervised Classification and  Regression, Unsupervised Clustering, Dimensionality Reduction, Interpretable Machine Learning, XGBoost, LightGBM, CatBoost, Random Forest, Local linear Forest,  GLM, SVM, K-Means, Prophet, Deep Learning Neural Networks,  LSTM,  RNN,  DeepAR, ARIMA, SARIMA, ETS, Ensemble, Hyperparameter tuning

Causal Machine Learning and Econometrics Skills:

Targeted Maximum Likelihood Estimation (TMLE), Causal Random Forest (CRF), Double Machine Learning (DML), Deep Instrumental Variables, Regression discontinuity design ,  Difference-in-Differences.

Econometrics (Bayesian) and Bayesian Machine Learning Skills:

Bayesian ARIMA, Bayesian structural times series (BSTS), SVAR, VAR, Bayesian VAR, Panel VAR, Stochastic Volatility BVAR, Time Varying BVAR, Dynamic factor model, MIDAS, Mixed Bayesian frequency.