// Curriculum Highlights
What You'll Learn
Linear & Logistic RegressionFrom OLS derivation to regularisation (Ridge, Lasso, ElasticNet)
Decision Trees & Random ForestsEntropy, information gain, bagging, feature importance
Gradient Boosting & XGBoostThe algorithm behind most Kaggle winners — built from intuition up
SVMs & Kernel TrickMaximum margin classifiers, RBF kernels, multi-class strategies
Clustering: K-Means & DBSCANUnsupervised segmentation for real customer & document datasets
Dimensionality Reduction (PCA)Eigendecomposition, explained variance, visualising high-d data
Model Selection & TuningCross-validation, GridSearchCV, RandomizedSearchCV, Optuna basics
Feature EngineeringEncoding, scaling, imputation, interaction terms, feature selection
Model Evaluation Deep DiveROC-AUC, precision-recall, confusion matrix, calibration
Pipelines with Scikit-learnProduction-ready ML workflows with Pipeline & ColumnTransformer
// 8 Algorithms You'll Master
From Classic to Cutting-Edge
Every algorithm is taught the same way: intuition first, maths second, code third, real dataset fourth. By the end you'll know when to use each one — and why.
Regression
Linear Regression
OLS, gradient descent derivation, Ridge & Lasso regularisation.
House prices · Salary predictionClassification
Logistic Regression
Sigmoid, log-loss, binary & multi-class, decision boundaries.
Churn · Fraud detectionTree-Based
Decision Trees
Gini impurity, entropy, pruning, overfitting control.
Medical diagnosis · Loan approvalEnsemble
Random Forest
Bagging, feature randomness, OOB error, feature importance.
Tabular data · Feature rankingBoosting
XGBoost / LightGBM
Sequential weak learners, learning rate, early stopping, SHAP.
Kaggle competitions · FinTechKernel Methods
Support Vector Machine
Maximum margin, kernel trick, soft margin, multi-class SVC.
Text classification · Image dataUnsupervised
K-Means Clustering
Inertia, elbow method, silhouette score, initialisation strategies.
Customer segmentationDimensionality
PCA
Eigendecomposition, explained variance ratio, scree plots, 2D viz.
Face recognition · EDA+
Naive Bayes, KNN, and more covered in bonus sessions
// What You'll Be Able to Do
Skills You'll Walk Away With
Proficiency Level by End of Course
Build & train supervised modelsExpert
Feature engineering & preprocessingAdvanced
Model evaluation & diagnosticsAdvanced
Hyperparameter tuningIntermediate
Kaggle & competitive MLIntermediate
Scikit-learn PipelinesAdvanced
// After This Course
Career Outcomes
Machine Learning Engineer
The core job title for this skillset. Building, testing, and deploying models at tech companies.
₹8–18 LPA fresher rangeData Scientist
DS roles at Indian startups and MNCs heavily overlap with ML Engineering. This covers both.
₹10–22 LPA fresher rangeKaggle Competitor
Enough tree-based and boosting knowledge to participate seriously in structured prediction competitions.
Portfolio builderResearch / Applied Scientist Intern
Top internship roles at Amazon, Google, and Microsoft expect end-to-end ML pipeline experience.
₹60–120k/month stipend// This course is for
🐍 Those who know Python & basic stats (or completed our first two courses)
🎓 B.Tech CSE/IT/ECE students who want to go beyond theory
💼 Software engineers pivoting into ML roles
❌ Not for: complete beginners with no Python experience
// Interview Intelligence
What Companies Actually Test
We analysed 200+ ML interview reports from Indian companies. Every topic in this course maps directly to what's being tested. Here's the breakdown by company type:
Product Startups (Zepto, Swiggy, Meesho)
- XGBoost end-to-end on tabular data
- Feature engineering from raw logs
- Evaluation: precision/recall tradeoffs
- Scikit-learn Pipelines for production
MNCs (Amazon, Microsoft, Flipkart)
- Explain bias-variance tradeoff
- How Random Forest reduces overfitting
- Cross-validation strategy choices
- ROC-AUC vs accuracy — when and why
AI/ML Startups (Sarvam, Krutrim, etc.)
- Build a model from scratch in NumPy
- PCA for dimensionality reduction
- Explain SHAP values intuitively
- Hyperparameter tuning with Optuna
Research Labs & PSUs
- Derive gradient descent step by step
- Difference between SVM kernels
- When to use clustering vs classification
- Regularisation: L1 vs L2 intuition
// Week by Week
Full Curriculum — 6 Weeks
- ML pipeline refresher: data → model → evaluation
- Linear Regression: OLS, gradient descent implementation
- Ridge, Lasso, ElasticNet regularisation — intuition + code
- Feature scaling: StandardScaler, MinMaxScaler, RobustScaler
- Encoding categorical variables: one-hot, ordinal, target encoding
- Train/validation/test split strategy & data leakage prevention
- Logistic Regression: sigmoid, log-loss, decision boundary
- Multi-class: OvR and softmax strategies
- Evaluation deep dive: accuracy, precision, recall, F1, ROC-AUC
- Imbalanced datasets: SMOTE, class weights, threshold tuning
- SVMs: maximum margin, support vectors, kernel trick
- RBF, polynomial kernels — when to use which
- Decision Trees: Gini impurity, entropy, information gain
- Depth, min_samples, pruning — controlling overfitting
- Bagging intuition: why averaging reduces variance
- Random Forest: feature randomness, OOB error
- Feature importance: MDI vs permutation importance
- ExtraTrees: when it beats Random Forest
- Gradient Boosting intuition: sequential residual fitting
- XGBoost: tree structure, regularisation, early stopping
- LightGBM: leaf-wise growth, categorical feature handling
- SHAP values: model interpretability for real interviews
- Hyperparameter tuning: GridSearchCV, RandomizedSearchCV, Optuna
- Kaggle walkthrough: top-5 submission on Titanic & House Prices
- K-Means: algorithm, inertia, elbow method, k-means++
- DBSCAN: density-based clustering, epsilon, min_samples
- Hierarchical clustering & dendrograms
- PCA: eigendecomposition, explained variance ratio, scree plot
- t-SNE for 2D visualisation of high-dimensional data
- Silhouette score, Davies-Bouldin for cluster evaluation
- Scikit-learn Pipeline & ColumnTransformer — production patterns
- Advanced feature engineering: polynomial, interaction, binning
- Feature selection: recursive elimination, SelectKBest, Boruta
- Model calibration: Platt scaling, isotonic regression
- Capstone: end-to-end ML system on a real business dataset
- Code review, portfolio submission, Silver badge issuance
// Hands-on Work
5 Portfolio-Quality Projects
PROJECT 01
Credit Default Prediction
Build a binary classifier on financial data. Handle severe class imbalance, engineer features from transaction history, and optimise for recall over accuracy.
PROJECT 02
E-commerce Customer Segmentation
Use K-Means and RFM analysis to segment 100k+ customers into actionable groups for marketing personalisation.
PROJECT 03
Kaggle: House Price Prediction
Full Kaggle submission pipeline — feature engineering, stacking models, hyperparameter tuning targeting top-10% leaderboard position.
PROJECT 04
Employee Attrition Predictor
A Random Forest pipeline on IBM HR data with SHAP explanations — exactly the kind of interpretable ML that enterprise analytics teams ask for in interviews.
🏆
Project 05 — Capstone: Open Business Problem
In week 6, you choose a real-world dataset and business problem. You scope, engineer, model, evaluate, and present a full ML system. Mentor reviews your code and presentation. This becomes your portfolio centrepiece — and the project you talk about in every interview.
// Your Credential
Silver Certificate Awarded
🥈
Newton JEE Silver Badge
ML Practitioner — Machine Learning Mastery
Appears on your LinkedIn profile
The Silver Badge — Your First Major Milestone
The Silver badge signals to recruiters that you can build real ML models end-to-end — not just run Scikit-learn demos. It's the first badge that directly unlocks ML Engineering and Data Science interview calls from mid-to-senior hiring managers.
1
Complete all 6 weeks & submit 5 projects2
Mentor approves your open capstone project3
Attend Silver badge presentation session4
Receive LinkedIn credential link within 48hrs5
One-click publish to LinkedIn under Certifications// Your Mentor
Meet Your Instructor
SK
Siddharth Kumar
Lead ML Engineer · Ex-Uber India & PhonePe
9 years building production ML systems — Uber's surge pricing models, PhonePe's fraud detection pipeline, and most recently leading ML at a Series B fintech. Siddharth is Newton JEE's most requested mentor. His teaching philosophy: every algorithm must be understood deeply enough that you can explain it to a non-technical stakeholder, debug it at 2am, and extend it to a new domain — in that order.
// Upcoming Batches
Pick Your Batch
Batch #18
Mar 22, 2025
Sat & Sun · 10:00–12:00 PM IST
2 seats left
Batch #19
Apr 12, 2025
Sat & Sun · 2:00–4:00 PM IST
10 seats open
Batch #20
Apr 26, 2025
Sat & Sun · 10:00–12:00 PM IST
15 seats open
// Ready to Start?
Enrol in This Course
₹7,999
₹14,000
Save ₹6,001 · 43% OFF
or ₹667/month · 12-month no-cost EMI
🔒 Secured by Razorpay · 100% refund after 2 sessions if unsatisfied
Everything included
✓ 18 live sessions (2hrs each) · 36 hrs total
✓ Lifetime recording access
✓ 5 projects incl. open capstone
✓ LinkedIn-verified Silver badge
✓ Model evaluation cheat-sheet
✓ Kaggle competition walkthroughs
✓ Mentor office hours (1hr/week)
✓ Resume & LinkedIn profile review
✓ Placement referral support
// Alumni Feedback
What Students Say
★★★★★
Siddharth's XGBoost session alone was worth the entire course fee. He showed us exactly how to tune it for a Kaggle dataset and I ended up in the top 8%. That result is on my resume now and every interviewer asks about it.
RT
Rahul Tripathi
B.Tech CSE → ML Engineer · Zepto
★★★★★
I had done two online ML courses before this. Neither of them taught me how to actually debug a model or explain predictions. The SHAP session and interview prep sections made Newton JEE a completely different tier of learning.
KP
Kavya Pillai
Software Eng → Data Scientist · Meesho
★★★★★
The open capstone project is where everything clicked. I built a loan default prediction system for an NBFC dataset, presented it to the mentor and batch, and that presentation turned into a portfolio piece that got me 4 interview calls in one week.
VN
Vikram Nair
Mech Engg → ML Engineer · PhonePe
★★★★★
The "What Companies Test" section they shared mid-course was eerily accurate. I had an Amazon interview the week after the SVM session and they asked almost exactly what Siddharth had warned us about. Cleared the ML round for the first time.
AS
Akshata Sawant
B.Tech IT → Applied Scientist · Amazon
// What's Next
Students Also Take