TDS Exam Engine
Tools in Data Science (TDS) solver for IIT Madras TDS community: step-by-step guides for GA0–GA8, ROE, Project 1 & 2 and end-term practice across T1, T2 and T3 2026.
Pick a term and exam, enter your IITM email, and open your workspace instantly in your browser.
What this portal can solve
19 exams, 3 terms and 525 deterministic solvers for IIT Madras Tools in Data Science — every solver runs in your browser from your IITM email.
Latest additions newest first · from T3 2026 and T2 2026
Newest solverGA 2 (API Engineering & Cloud Services)T3 2026Current term15 solvers
Statistics APIs, CORS, OAuth JWKS verification, Redis counters, Prometheus observability, and LLM endpoints.
- Q1: HTTPBin Terminal Lab (termlog)
- Q2: Differentiate Model Intelligence — Automatic Evaluation
- Q3: Deploy a CORS-Aware Metrics API
- Q4: OAuth 2.0 / OIDC Token Verification Service
- Q5: Resolve 12-Factor Config Precedence
- Q6: Multi-Container App via Compose, Exposed by Tunnel
- Q7: Deploy a POST Analytics Endpoint
- Q8: Production Observability: Metrics, Health & Structured Logs
- Q9: Expose a Local LLM through a Tunnel
- Q10: Local LLM Structured-Output Service
- Q11: API Engineering: Idempotency + Pagination + Rate Limit
- Q12: Middleware Stack: Rate-Limit + CORS + Request Context
- Q13: Recover a Fragmented LLM Stream
- Q14: Diagnose Docker Build Cache Misses
- Q15: Create a Custom Ollama Model (termlog)
GA 1 (Developer Tools)T3 2026Current term16 solvers
Developer tooling, API servers, Git forensics and an agent ledger — including live-API questions.
- Q1: VS Code Version
- Q2: Find the Code in the Response Headers
- Q3: POST HTTP Requests with uv
- Q4: Run Prettier with npx
- Q5: Fix Broken JSON File
- Q6: Resolve 12-Factor Config Precedence
- Q7: Reorganize Files with Shell Commands
- Q8: Extract and Analyze CSV Sales Data from ZIP
- Q9: Host Your Portfolio on GitHub Pages
- Q10: Git Forensics — Recover the Leaked Token
- Q11: Cosine Similarity Search
- Q12: One Prompt, Two Models, Two Answers
- Q13: Decode It in the Terminal (termlog)
- Q14: Build a Live MCP Server
- Q15: The Ledger Agent
- Q16: Geometric Reasoning — Tangram
GA 8 (MLOps & LLM Systems Gateway)T2 2026Recent term10 solvers
MLOps, training-data integrity and LLM gateway systems.
- Q1: Immutable Training Corpus
- Q2: Leakage-Safe BigQuery ML Gate
- Q3: MLflow Evidence Promotion Gate
- Q4: PEFT Choice & Training Repair
- Q5: Quantized Model Admission Gate
- Q6: Content-Addressed DAG Pipeline
- Q7: Verifiable Model Bundle Verifier
- Q8: Per-Layer QLoRA Adapter Synthesis & Parameter Audit
- Q9: PyTorch Training Loop & MLflow Fingerprint
- Q10: Green AI & Model Card Carbon Frontmatter
GA 7 (Policy Gates & OSINT)T2 2026Recent term10 solvers
CI/CD release gates, policy checks and OSINT.
- Q1: CI/CD Container Release Gate
- Q2: LLM Action Firewall
- Q3: Terraform Plan Policy Gate
- Q4: LLM Output Handling Gate (LLM05)
- Q5: OSINT Corroboration Engine
- Q6: Street View OSINT: Where Is This?
- Q7: Advanced Search Operators
- Q8: WAF Rule Order
- Q9: Media Forensics
- Q10: Audit a GitHub Actions Workflow
Browse all 19 exams and 525 solvers
T2 2026 May–Sep 2026
GA 6 (Data Forensics & Automation)T2 2026Recent term10 solvers
Image and data forensics with automation scripts.
- Q1: Image Forensics — Recover a Rotated and Mirrored Grid
- Q2: The Multi-Model Robustness Audit
- Q3: DuckDB — Multi-Table Linear Regression Analysis
- Q4: Playwright — Reconcile a Paginated Shadow-DOM Incident Audit
- Q5: DuckDB — Reconcile a Versioned Nested-JSON Invoice Ledger
- Q6: Crawl a Static Site Respecting robots.txt
- Q7: Scrape Books to Scrape by Category and Value
- Q8: GitHub Action — Scrape Table Sums with Playwright
- Q9: Sum Table Values with Playwright
- Q10: Decode a Hidden Modem Signal in Audio
GA 5 (Agentic Systems Safety)T2 2026Recent term11 solvers
Agent safety, sandboxing and guard-rail challenges.
- Q1: Solve a Generated Maze Offline
- Q2: Spec-Driven Development: The Proration Bug
- Q3: Agent Harness — Pre-Tool-Call Guardrail Hook
- Q4: Skill Safety Audit — Scanner API
- Q5: Agent Harness — Run Budget & Loop Guard
- Q6: Build a Live MCP Server
- Q7: Prove You Contained It (LXD Sandbox)
- Q8: Guardrail Red-Team Round-Trip
- Q9: Lethal-Trifecta Mailroom Action Gate
- Q10: A2A Invoice Action Agent
- Q11: Build an Observable Incident-Response Agent
GA 4 (RAG & Vector Search)T2 2026Recent term13 solvers
Chunking, embeddings, hybrid retrieval and vector search.
- Q1: End-to-End Chunking & Hybrid Retrieval Pipeline
- Q2: RAG Evaluation Harness (RAGAS-style Metrics)
- Q3: Anti-Hallucination Grounded Answer API
- Q4: Production Vector Search API with Re-ranking
- Q5: GraphRAG Pipeline: Extract → Index → Query
- Late Chunking with Contextual Retrieval
- Semantic Caching and Query Augmentation
- Multimodal Embedding Calibration
- HyDE: Hypothetical Document Embeddings for Query Augmentation
- ANN Index: IVF Recall vs. Latency Tradeoff
- Semantic Deduplication with Numeric-Fact Guardrails
- Context Assembly: Beating Lost-in-the-Middle
- Reciprocal Rank Fusion by Hand
GA 3 (System & API Architecture)T2 2026Recent term13 solvers
Pipelines and API architecture, including LLM-backed services.
- Q1: Automated Video Curation Pipeline
- Q2: Multimodal Image Question-Answering API
- Q3: Fixed Schema Invoice Extraction API
- Q4: Dynamic Schema Structured Extraction API
- Q5: Cosine Similarity Search
- Q6: Korean Audio Dataset API Verification
- Q7: Invoice Intelligence — Structured Extraction API
- Q8: Semantic Search — Top-K Ranking API
- Q9: Reliable Reasoning — Word-Problem Solver API
- Q10: Proof-of-Work Nonce Hunt
- Q11: Context Window Heist
- Q12: Spin Up the CLI — LLM CLI Tools
- Q13: Embedding Trapdoors — Nearest Neighbor Search
GA 2 (API Engineering & Cloud Services)T2 2026Recent term10 solvers
Statistics APIs, CORS, serverless endpoints and cloud services.
- Q1: Descriptive Statistics with strict CORS
- Q2: OAuth JWT Verification Service
- Q3: 12-Factor Config Precedence
- Q4: Multi-Container Redis Stack with Tunnel
- Q5: POST Analytics Endpoint
- Q6: Production Observability (Metrics & Logs)
- Q7: Expose a Local LLM through a Tunnel
- Q8: Local LLM Structured-Output Service
- Q9: API Engineering: Idempotency + Pagination + Rate Limit
- Q10: Middleware Stack (Rate-Limit, CORS & Context)
GA 1 (Developer Tools)T2 2026Recent term20 solvers
Git, GitHub Actions, JSON, shell tools and developer workflows.
- Q1: VS Code Version (code -s)
- Q2: Multi-cursor → JSON (jsonhash)
- Q3: Prettier SHA256 of README.md
- Q4: LLM Bash Pipeline
- Q5: Reorganize Files with Shell Commands
- Q6: Record Terminal Session with asciinema
- Q7: Three-Way Configuration Merge: Detect Conflicts
- Q8: Git Time Travel: History Investigation
- Q9: GitHub Pages with Email
- Q10: GitHub Action with Email in Step Name
- Q11: POST HTTP requests with uv
- Q12: Use DevTools
- Q13: Fix Broken JSON File
- Q14: Extract JSON from a ZIP
- Q15: Unicode → Markdown Converter
- Q16: Infer SQL Schema from CSVs
- Q17: SQL: Average Order Value
- Q18: Markdown Deployment Architecture
- Q19: Debug and Improve a Failing Prompt
- Q20: Verify and Fix AI-Generated Code
GA 0 (Warm-up Exam)T2 2026Recent term25 solvers
Warm-up exam: charts, data repair and file tasks.
- Q1: Axis Scale Manipulation Repair
- Q2: Binary Evaluation Rubric
- Q3: The Bug Hunter (Property-Based Testing)
- Q4: Quality Control (Sample Variance)
- Q5: Code Interpreter Service
- Q6: Color Encoding Mismatch Repair
- Q7: Crawl HTML: Link Normalization
- Q8: CSS Selector Discount Sum
- Q9: dbt Operations performance mart
- Q10: FastAPI Students CSV Service
- Q11: FastAPI Batch Sentiment Analysis
- Q12: Get LLM to say YES (Jailbreak)
- Q13: GitHub Action Workflow
- Q14: Image Grayscale Reconstruction
- Q15: LLM Sentiment Analysis (httpx)
- Q16: Move and rename files
- Q17: Network Game: Graph Detective
- Q18: Local Ollama Endpoint
- Q19: Replace Across Files
- Q20: Sort and Filter JSON Catalog
- Q21: Average Salary by Department
- Q22: Process Unicode Files
- Q23: Use DevTools
- Q24: Use GitHub
- Q25: Vercel Latency
Project 2T2 2026Recent term8 solvers
Data case studies solved with analysis and write-ups.
- Case Study 1A — DTH Month-End Mystery
- Case Study 1B — DTH Complaints Went Quiet
- Case Study 2A — Solar Inverter Smell Test
- Case Study 2B — Solar 31.6% Impact Claim
- Case Study 3A — Swiss Mismatch Control
- Case Study 4A — QC Queue Smell Test
- Case Study 3B — Is the Irish Preference Claim Supported?
- Case Study 4B — Spare-Parts Search
Project 1T2 2026Recent term5 solvers
AI problem discovery and requirements-interview tasks.
- Q1: AI Problem Discovery — Requirements Interview
- Q2: Differentiating Model Intelligence
- Q3: AI Agent GCS Bucket Setup
- Q4: AI Agent Dataset Upload to GCS
- Q5: Data Analyst Telegram Bot
ROE Re-ExamT2 2026Recent term12 solvers
Re-examination questions: OSINT, forensics, APIs and data tasks.
- Q4: Street View OSINT — Where Is This?
- Q6: Donate Your Marks
- Q7: Donate Your Marks — Tell Them Why (Audio)
- Q1: Incident Atlas — Georegister and Route
- Q2: Unicode Doppelganger Ledger Forensics
- Q3: HTTP Cache Time Machine — Reconstruct the Shared Cache
- Q10: Prompt for Detecting AI-Generated Content
- Q11: AI Opportunity Discovery
- Q9: The Unusual Useful Essay
- Q8: Something You Did On Your Own Initiative (Audio)
- Q12: Insights from NSS80 Telecom Data
- Q13: AI Tutor Challenge
End-Term Mock (175 MCQ + 25 MSQ + 100 Subjective)T2 2026Recent term300 solvers
A 300-question practice bank (MCQ, MSQ and subjective) with worked explanations.
300 practice questions with explanations — open the workspace to browse and practise them.
T1 2026 Jan–Apr 2026
GA 8 (MLOps & DevOps)T1 2026Archive15 solvers
GitHub Actions, hash chains and MLOps tasks.
- GitHub Actions: Secret Hash Chain
- GCP Gemini API: Math Puzzle
- Deploy a FastAPI Iris Classifier
- Hugging Face Spaces: ML Model API
- Docker Multi-stage Build: Verify Your Image
- MLOps Bash Script: Deterministic Output
- Pre-commit Hooks + CI Gate: ruff
- MLOps Concepts: Hash-Verified Quiz
- GCP Cloud Run: Deploy a Compute Service
- GCP Cloud Functions: HTTP Text Processor
- GCP AI Studio: Gemini Text Classification
- GCP Cloud Run: Deploy an ML Classifier
- GCP Cloud Run: Environment Variable Configuration
- GCP Cloud Run: Hash Verification API
- GCP AI Studio: Gemini JSON Data Extraction
GA 7 (Data Visualization)T1 2026Archive15 solvers
Chart repair, colour encoding and visualization fixes.
- Fix the Color Encoding Mismatch
- Chartjunk Removal and Data-Ink Ratio Repair
- Narrative Integration Repair
- Data-Narrative Number Reconciliation
- Broken Aggregation and Sort Order Repair
- Scale Manipulation Repair
- Headline Rewriting: Description vs Finding
- Prompt Reverse-Engineering
- Presentation Prompt Structural Repair
- Ranked Anomaly Detection
- Poisoned Document Detection
- Flaw Priority Ranking
- Chart Error Detection
- Cloud Deployment Cost & Performance Analysis
- Service Latency Spike Detection
Project 2 Part BT1 2026Archive2 solvers
Blockchain forensics and QR tracing for Project 2 Part B.
- Q3: QR Forensics — Solana Devnet Tracer
- Q4: Discourse KB Solver (50 Tasks)
ROE Re-ExamT1 2026Archive15 solvers
Re-examination questions across APIs, encodings and data tasks.
- Collaborative Token Exchange
- 韓国語音声データセットAPI検証
- Regex Golf Challenge
- Maze Solver with Constraints
- Cipher Trail (Caesar Decode)
- Layered Encoding Challenge
- Region Containing Point (Geospatial)
- Reorganize Files (Shell SHA256)
- Refactor Python Code (PEP 8)
- Fix Broken JSON
- Cross-Lingual Entity Disambiguation
- Trick Question (Hidden Chinese)
- Terminal Recording (asciinema)
- FastAPI IoT Sensor Analytics
- Video Attendee Extraction (Gemini)
1. Purely Educational Simulation Sandbox
SANDBOX CODEThis workspace serves strictly as a client-side offline simulation sandbox for personal study, algorithm verification, and deep-dive technical research. All solvers execute locally in your browser's sandboxed context with zero network transmissions, API telemetry, database syncs, or communication with official IIT Madras grading frameworks.
2. Malpractice Prohibition & Honor Code
RESTRICTEDUse of this application as a cheating helper, live exam bypass, or learning proxy during active graded evaluations is strictly prohibited. By accessing these resources, you covenant to remain bound by Section 4.1 of the IIT Madras Student Code of Conduct (Academic Integrity). You pledge to utilize this system solely to verify manual derivations, study computational patterns, and check your work. Circumventing active learning violates university regulations.
3. Non-Affiliation & Fair Use Exemption
FAIR USEThis project is a entirely independent open-source workspace. It is NOT affiliated, sponsored, endorsed, or associated with the Indian Institute of Technology Madras (IIT Madras) or the Tools in Data Science (TDS) course administration. The mathematical algorithms, educational templates, and structural cleanroom code are published under the Fair Use Doctrine (Section 52(1)(a) of the Indian Copyright Act, 1957) strictly for private study and research exemption.
4. Liability Covenant & Takedown Protocol
COOPERATIVEThe developers, contributors, and the GyaanFlow team accept zero liability or legal responsibility for individual user conduct or academic standings. We operate in complete transparency. If any representative, coordinator, or instructor from the IITM TDS course administration has compliance concerns or takedown requests, please connect with us via our verified LinkedIn Profile. We pledge immediate cooperation and swift action.