Topics Roadmap 90-Day Track War Room

Weekend War Room

Every weekend: revision checklist, quiz blocks, and one-by-one mini viva with ideal answers and follow-ups.

Database Links

users(id, name, email)
war_room_sessions(id, user_id, weekend_date, score)
war_room_answers(id, session_id, question_id, answer_text, rubric_score)

Procedural Links

POST /api/war-room/session/start
POST /api/war-room/answer
POST /api/war-room/quiz/submit
GET /api/war-room/session/{id}/report

Scoring Rules

MCQ: objective auto-score
Viva: rubric-based conceptual scoring
Weekend rank: weighted aggregate of quiz + viva + coding output

Weekend Revision Checklist

Explain linear regression without writing formulas.Explain why MSE penalizes outliers more than MAE.Derive gradient descent intuition using a bowl-shaped loss surface.Identify overfitting from train and validation curves.Choose between linear regression and random forest for a noisy tabular dataset.Explain one concrete data leakage example from feature engineering.Explain why regularization can increase train loss but improve test metrics.Define precision and recall using a fraud detection example.Explain when accuracy is a misleading metric.Explain why feature scaling matters for KNN and SVM.Give one case where decision trees fail and why.Explain how random forests reduce variance.Explain the bias-variance tradeoff with model complexity changes.Give one practical way to detect data drift in production.Explain one strategy for model selection under latency constraints.Explain why baseline models are mandatory before advanced models.Explain how you would debug poor minority-class recall.Describe one post-deployment model failure case and mitigation.

ML Atlas

A premium ML knowledge base. Built for engineers, by engineers.