Machine Learning Systems
End-to-end ML pipelines — primarily in R with tidymodels, plus Python with scikit-learn. From feature engineering and model selection to deployment and monitoring.
Hi, I'm
Forhad
Data Science & Machine Learning Instructor
I teach professionals and learners how to build production-grade ML systems — from demand forecasting with ARIMA & XGBoost to interactive Shiny dashboards and reproducible Quarto reports. R-first, Python-capable. Real datasets. No fluff.
10+
Courses
5K+
Learners
2
Languages
# Build. Visualize. Predict.
library(tidymodels)
library(modeltime)
model_spec <- boost_tree(
trees = 500,
learn_rate = 0.01
) %>%
set_engine("xgboost") %>%
set_mode("regression")
forecast <- model_spec %>%
fit(value ~ ., data = train)
# → RMSE: 0.042 ✓
About Me
Building clarity in complex systems
I'm a data science instructor focused on bridging the gap between theoretical ML and production-ready analytics — built primarily with R and the tidyverse ecosystem, complemented by Python where it shines. My courses are designed for professionals who want to solve real business problems, not just pass exams.
With experience across demand forecasting, hierarchical time-series modeling, interactive Shiny applications, reproducible Quarto-based reporting, and Python-based ML pipelines with scikit-learn and LightGBM, I bring structure and clarity to topics that are often taught in a disorganized way. Every course I build uses real-world datasets, production patterns, and step-by-step workflows.
I produce courses in both English (for a global audience) and Bangla (for learners in Bangladesh and the Bengali-speaking community), because quality technical education should be accessible in your own language.
Structured Curriculum
Clear learning paths from fundamentals to production
Real Datasets
Industry-relevant data, not toy examples
Bilingual Content
Courses in English & Bangla for wider reach
Core Tech Stack
Expertise
What I teach & build
Deep focus areas with real production experience
Forecasting & Time Series
Demand forecasting, sales prediction, and anomaly detection using R's fable & modeltime and Python's statsmodels & LightGBM on real business datasets.
Hierarchical Modeling
Reconciled forecasts across product hierarchies, geographies, and business units — ensuring coherent predictions at every level.
MLOps & Kubernetes
Containerized model serving with R plumber and Python FastAPI, CI/CD for ML, orchestration with Kubernetes, and experiment tracking.
Shiny Apps & Quarto
Production-grade interactive dashboards with Shiny, and reproducible scientific reports, blogs, and books with Quarto and R Markdown.
Business Analytics Strategy
Data-driven decision frameworks, KPI design, dashboard strategy, and translating analytical outputs into actionable business insights.
Courses
Learn at your own pace
Structured, practical courses in English and Bangla
Demand Forecasting with R & modeltime
Build production-grade demand forecasting systems using tidymodels, modeltime, and XGBoost on real retail datasets.
Coming SoonEnd-to-End ML Pipelines with tidymodels
From raw data to deployed model — learn to architect, build, and maintain ML systems in R that scale with recipes, parsnip, and plumber.
Coming SoonProduction Shiny Apps & Quarto Reports
Build interactive dashboards with Shiny, deploy them at scale, and create reproducible reports and books with Quarto.
Coming SoonMLOps for R & Python: Docker, APIs & K8s
Containerize R plumber and Python FastAPI models, set up CI/CD pipelines, and deploy on Kubernetes for production ML.
Coming SoonPython ML Essentials: scikit-learn & LightGBM
A focused Python track — learn scikit-learn pipelines, LightGBM tuning, and Pandas-based feature engineering for production ML.
Coming SoonR দিয়ে ডেটা সায়েন্স ফান্ডামেন্টালস
R, tidyverse, ggplot2 এবং বাস্তব ডেটাসেট ব্যবহার করে ডেটা সায়েন্সের মৌলিক ধারণা শিখুন।
শীঘ্রই আসছেমেশিন লার্নিং A-Z বাংলায় (R দিয়ে)
tidymodels দিয়ে সুপারভাইজড থেকে আনসুপারভাইজড লার্নিং — সম্পূর্ণ বাংলায় হাতে-কলমে শিখুন।
শীঘ্রই আসছেShiny ড্যাশবোর্ড ও Quarto রিপোর্ট বাংলায়
ইন্টারেক্টিভ Shiny অ্যাপ তৈরি এবং Quarto দিয়ে রিপ্রোডিউসিবল রিপোর্ট ও ডকুমেন্টেশন শিখুন।
শীঘ্রই আসছেPython দিয়ে মেশিন লার্নিং বাংলায়
scikit-learn, Pandas এবং LightGBM ব্যবহার করে হাতে-কলমে মেশিন লার্নিং শিখুন — সম্পূর্ণ বাংলায়।
শীঘ্রই আসছে
Blog
Latest articles
Insights on ML systems, forecasting, and data-driven decisions
Why XGBoost in R Outperforms ARIMA for Retail Demand Forecasting
A practical comparison using modeltime and real retail data — when to use statistical methods vs. gradient-boosted trees in the tidymodels framework.
Read article →Building Production Shiny Dashboards with golem & Docker
From prototype to production — structuring Shiny apps with golem, containerizing with Docker, and deploying to ShinyProxy or Posit Connect.
Read article →R vs Python for ML Pipelines: When to Use What
A practical guide to choosing between tidymodels and scikit-learn — comparing workflows, deployment options, and where each language truly excels.
Read article →Hierarchical Forecasting in R with fable
How to build coherent hierarchical forecasts using fable and reconciliation methods — from store-level to national-level, all adding up correctly.
Read article →Quarto for Data Scientists: From Notebooks to Websites
A hands-on guide to using Quarto for reproducible reports, presentations, blogs, and full websites — the modern replacement for R Markdown.
Read article →Feature Engineering for Time Series in R with timetk
A recipe-based approach to creating calendar features, lags, rolling windows, and Fourier terms for ML-based time series forecasting in R.
Read article →
Contact
Get in touch
Have a question or want to collaborate? Reach out anytime.