2hr/day Program · Mon–Fri · 3 months · 120 hrs
Data Science
Core Track
Core Track
Pure manual analysis and modelling. No AI shortcuts. Python + Statistics + Machine Learning + SQL + Visualization + Model Deployment. Students graduate able to take raw data, build predictive models, and communicate findings — the exact skill junior data scientist interviews test. Placement support included.
2hr Core Track — 120 hrs
Python + NumPy + Pandas Statistics + Math SQL ML: Regression + Classification Clustering + NLP basics Streamlit deployment Placement support8hr Advanced Track — 768 hrs adds
Deep Learning · TF/Keras CNN + LSTM + Transformers HuggingFace NLP Time Series · Prophet MLflow · Model tracking AI fundamentals + Gemini API ML paradigms + DFS/BFS Internship · Placement supportFoundation — Python for Data Science
P1
Python Fundamentals for Data Science
Variables, data types, lists, dicts, loops, functions, file handling, error handling. Taught through data tasks: parse a CSV, count categories, filter records. Jupyter Notebook workflow. Students feel Python as a data tool from Day 1 — not abstract coding.
Python 3Jupyter
P2
NumPy — Numerical Computing
Arrays vs lists, creating, reshaping, slicing, boolean indexing, vectorized math, statistical functions (mean, std, percentile). Why data scientists need arrays: speed on large datasets and the mathematical foundation under every ML library.
NumPyArrays
P3
Pandas — Data Wrangling Power Tool
Series and DataFrames. Reading CSV, Excel, JSON. Selecting, filtering, sorting. Handling nulls, duplicates, type conversion, string cleaning. groupby aggregations, merge/join, datetime features. Real task: clean a 10,000-row messy bookings dataset — the daily work of a data scientist before any model touches data.
▸ Milestone: raw messy file → clean, analysis-ready dataset → summary statistics.
PandasgroupbyData cleaning
P4
Data Visualization — Matplotlib + Seaborn
Chart selection logic: when bar vs line vs scatter vs histogram vs heatmap. Matplotlib figures, labels, styling, subplots. Seaborn for statistical plots: distributions, heatmaps, pairplots. One rule from Day 1: every chart must answer a specific business question.
MatplotlibSeaborn
Mathematics for Data Science
M1
Statistics — The Core of Data Science
Descriptive stats: mean, median, mode, variance, std, percentiles, skewness, outliers (IQR + Z-score). Probability basics: events, conditional probability, Bayes intuition. Distributions: normal, binomial, Poisson. Central Limit Theorem. Confidence intervals. Hypothesis testing: p-value, Type I/II errors, t-test, chi-squared. A/B testing logic. Every concept tied to a business decision.
StatisticsHypothesis testingProbability
M2
Linear Algebra + Gradient Descent Intuition
Vectors and matrices in data terms (rows = records, columns = features). Matrix multiplication, dot product, transpose. Gradient descent — what it is and why ML models use it to learn. Loss function intuition. Taught in 2 sessions — just enough to understand ML model internals without fear.
Linear AlgebraGradient descent
SQL for Data Scientists
S1
SQL — Data Extraction and Aggregation
SELECT, WHERE, ORDER BY, GROUP BY, HAVING, all JOIN types, subqueries, CASE, date functions, string functions. 30 practice queries on a realistic dataset. SQL rounds decide most data science interviews — drilled hard.
MySQLJoinsAggregation
Machine Learning — Core Track
ML1
ML Workflow — End-to-End Pipeline
The standard ML pipeline: define problem → collect data → clean → EDA → feature engineering → model → evaluate → improve → deploy. Overfitting vs underfitting. Train/validation/test split. Cross-validation. This module gives students the mental map before any algorithm is taught.
ML pipelinescikit-learn
ML2
Supervised Learning — Regression
Linear Regression: cost function, gradient descent, R², MAE, RMSE. Polynomial Regression. Ridge and Lasso regularization. Real task: predict service booking price from provider data. Students write the model, evaluate it, and explain coefficients in business language.
▸ Mini project: booking price predictor using Linear Regression.
Linear RegressionRidge/Lasso
ML3
Supervised Learning — Classification
Logistic Regression, Decision Trees, Random Forest, k-Nearest Neighbors. Evaluation: accuracy, precision, recall, F1, ROC-AUC, confusion matrix. Feature importance. Real task: predict booking cancellation (Yes/No). Students understand which metric to use when — recall for healthcare, precision for spam.
ClassificationRandom ForestROC-AUC
ML4
Unsupervised Learning — Clustering + PCA
k-Means: centroid logic, inertia, elbow method. Hierarchical clustering concept. PCA: dimensionality reduction — visualize high-dimensional data in 2D. Real task: customer segmentation on booking data — group providers into performance tiers without labels.
k-MeansPCA
ML5
Feature Engineering + Model Tuning
Encoding categorical variables (one-hot, label, target encoding). Feature scaling (StandardScaler, MinMaxScaler). Handling imbalanced classes. Hyperparameter tuning with GridSearchCV. scikit-learn Pipeline objects. Students tune their classification model and document improvements.
Feature engineeringGridSearchCVPipeline
ML6
NLP Basics — Text as Data
Tokenization, stop word removal, stemming, lemmatization. Bag-of-Words, TF-IDF. Sentiment analysis with scikit-learn. Real task: classify Zuvio customer reviews as positive/negative. Students see how text becomes numbers ML can process.
NLPTF-IDFSentiment analysis
Deployment + Portfolio
D1
Streamlit — Turn Models into Live Apps
Build a web UI for a trained model using Streamlit. Widgets, dataframe display, charts, model prediction output. Real task: wrap the cancellation predictor in a Streamlit app — user enters booking details, app returns cancel risk. Deployed to Streamlit Community Cloud for a public portfolio URL.
▸ Deliverable: Live Streamlit ML app — public URL for resume and LinkedIn.
StreamlitStreamlit Cloud
G1
Git + GitHub — Data Science Portfolio
Init, add, commit, push, branches, .gitignore (model files, data), README with project story. Notebooks pushed from early on. A visible GitHub portfolio — recruiters and hiring managers check this before the interview.
GitGitHub
Capstone
C1
DataSpark — End-to-End Data Science Project
Full pipeline on the Zuvio marketplace dataset: SQL extraction → Pandas cleaning → statistical analysis → EDA with charts → feature engineering → cancellation classifier → price predictor → NLP review sentiment → Streamlit app → written insight report with 5 business recommendations. GitHub with README. Resume + LinkedIn setup. Top 40 interview Q&A from real data science drives. Mock technical + HR round. Placement support through Appin network.
▸ Deliverable: End-to-end DS project on GitHub + live Streamlit app URL + insight report.
Program Fee
₹29,899
2hr/day · 3 months · 120 hrs · Placement support included
Job Descriptions — 2hr Graduates
Real JDs from Naukri, Internshala, Glassdoor, Indeed India — April 2026. Click any card to expand.
Junior Data Scientist — Fresher
Futops Technologies · Pune / Chennai · ₹2.5–4 LPA
Role
Work alongside senior data scientists to extract insights from data. Build ML models, clean datasets, support business decisions. End-to-end data pipeline work.
Required Skills
PythonPandasNumPyscikit-learnSQLStatisticsGit
Interview
Round 1: Python + Statistics + ML written test. Round 2: Walk through your project. Round 3: HR.
Search similar on Internshala →Data Scientist Intern → Full-Time
Analytics firms · Chennai / Remote · ₹15K/month → ₹3–4 LPA
Role
Data cleaning pipelines in Pandas, building classification models, maintaining dashboards, writing automated reports. 6-month internship with full-time conversion.
Required Skills
PythonPandasML basicsSQLStatisticsGit
Search similar on Internshala →ML Engineer Trainee
IT product startups · Bangalore / Chennai · ₹3–4.5 LPA
Role
Build and evaluate ML models. Feature engineering, model training with scikit-learn, evaluation metrics. Deploy models as Streamlit apps.
Required Skills
Pythonscikit-learnFeature engineeringSQLStreamlitGit
Search similar on Indeed →Data Analyst + ML — Fresher
IT analytics teams · Chennai / Coimbatore · ₹2.5–3.5 LPA
Role
Assist in data collection, analysis, visualization, and model implementation. Develop statistical analyses and collaborate with cross-functional teams on insights.
Required Skills
PythonPandasSQLData visualizationML basics
Search similar on Internshala →NLP / Text Analytics Junior Role
AI-enabled IT firms · Chennai / Remote · ₹3–4.5 LPA
Role
Text classification, sentiment analysis, entity extraction. Preprocessing pipelines, TF-IDF, basic model building for NLP client projects.
Required Skills
PythonNLPTF-IDFscikit-learnPandasGit
Search similar on Indeed →Junior Data Scientist / AI Engineer
SymphonyAI / mid-size AI companies · Coimbatore / Chennai · ₹3–5 LPA
Role
End-to-end data and AI projects — data collection, preprocessing, model development, evaluation. Work with product and engineering teams on AI-driven features.
Required Skills
PythonStatisticsMachine learningSQLNLP basicsGit
Search on Glassdoor →Research Analyst / Junior Data Scientist — Walk-In
Cyber Infrastructure Inc. (CIS) · Chennai / Coimbatore · ₹2.5–3.5 LPA
Role
Statistical analysis, basic ML model building, report writing, data preprocessing. Walk-in drive format — Python + stats written test, then project walkthrough.
Required Skills
PythonStatisticsSQLPandasProblem solving
View JD on CIS →Data Science Fresher — E-commerce / Marketplace
E-commerce companies · Bangalore / Remote · ₹3–5 LPA
Role
Funnel analysis, customer segmentation, prediction models for conversion and churn — same domain as the DataSpark capstone project.
Required Skills
PythonPandasSQLClusteringClassificationStreamlit
Search on Naukri →2hr Program Summary
2 Projects
Mini ML model + DataSpark capstone with live Streamlit URL
Full DS stack
Python+NumPy+Pandas+Stats+ML+NLP+SQL+Streamlit
3 months
Mon–Fri · 2hrs/day · 120 hrs
Placement
Resume + mock interview + placement support
8hr/day Program · Mon–Sat · 4 months · 768 hrs
Data Science
Advanced Track
Advanced Track
Everything in the 2hr Core Track plus Deep Learning, TensorFlow/Keras, CNN, LSTM, advanced NLP with HuggingFace Transformers, Time Series forecasting, MLflow model tracking, AI fundamentals, prompt engineering, Gemini API integration, GitHub Copilot workflow, a live deployed ML app, and a real internship. Placement support included for both programs.
Core Track — Months 1–3
2hr Core Track content — all topics covered in months 1–3
Python + NumPy + Pandas + Visualization + Statistics + Math intuition + SQL + ML core (Regression, Classification, Clustering, NLP basics, Feature Engineering) + Streamlit + Git. Same curriculum as the 2hr Core Track, completed in the first 3 months of the 4-month program.
Advanced Add-ons — Month 4 Onwards
Deep Learning — Neural Networks
DL1
Neural Network Fundamentals
What a neuron is. Layers (input, hidden, output). Activation functions: ReLU, Sigmoid, Softmax. Forward propagation. Backpropagation and gradient descent — taught with diagrams, not math exams. Vanishing gradient problem. Why deep networks outperform shallow ones on complex patterns.
Neural NetworksBackpropagation
DL2
TensorFlow + Keras — Building Neural Nets
TensorFlow 2.x. Keras sequential and functional API. Build, compile, train, evaluate. Dropout and BatchNormalization. Callbacks: EarlyStopping, ModelCheckpoint. Real task: tabular classification using a neural network on Zuvio bookings — compare against Random Forest baseline to see when deep learning helps.
▸ Mini project: Neural network classifier — compared with Random Forest baseline.
TensorFlowKeras
DL3
CNN — Image Classification with Transfer Learning
What Convolutional layers do. Pooling. CNN architecture. Transfer learning with MobileNetV2 — pre-trained weights, fine-tuning. Real task: classify service category images. Students understand why CNNs power vision AI without implementing from scratch.
CNNTransfer learningMobileNetV2
DL4
RNN + LSTM — Sequence and Time Data
Why sequence order matters. Recurrent connections. LSTM cell: input, forget, output gates in plain English. Real task: predict next month's booking volume using LSTM on historical data — time series forecasting.
RNNLSTMTime series
Advanced NLP + Transformers
NLP1
Word Embeddings — Word2Vec + GloVe
Why bag-of-words fails for meaning. Word embeddings: words as vectors, similarity = proximity. Word2Vec intuition. Using pre-trained embeddings to improve the review sentiment classifier from the Core Track.
Word2VecEmbeddings
NLP2
HuggingFace Transformers — Applied
What the attention mechanism does. HuggingFace pipeline API — zero-shot classification, NER, summarization in 5 lines of Python. Fine-tuning a pre-trained model on a small custom dataset. Real task: classify Zuvio support tickets using zero-shot model — compare with TF-IDF baseline.
▸ Applied: HuggingFace transformer for ticket classification vs TF-IDF baseline comparison.
HuggingFaceTransformersBERT
Time Series Forecasting
TS1
Time Series Analysis + Forecasting
Trend, seasonality, noise. Stationarity and ADF test. ARIMA explained simply. Facebook Prophet: business-friendly forecasting, holiday effects, uncertainty intervals. Comparing Prophet vs LSTM on the same dataset. Real task: forecast next 4 weeks of Zuvio service bookings.
▸ Mini project: Booking volume forecaster with Prophet + LSTM comparison.
ARIMAProphetForecasting
MLOps Basics — Model Lifecycle
OPS1
MLflow — Experiment Tracking
What MLflow is and why production teams use it. Logging parameters, metrics, models. Comparing experiment runs. Model registry concept. Students track all capstone experiments in MLflow — makes their portfolio look senior-level to hiring managers used to seeing untracked notebooks.
MLflowExperiment tracking
OPS2
Advanced Streamlit + Flask Model API
Wrap a trained model in a Flask REST API — send JSON input, get prediction JSON. Advanced Streamlit: file upload, multi-page app, caching, session state. Build the DataSpark Pro app with Gemini AI insight summaries. Deploy to Streamlit Community Cloud.
▸ Deliverable: Live DataSpark Pro app — upload data → ML predictions → AI insights → deployed URL.
Flask APIStreamlit Pro
AI + ML Fundamentals — 8hr Exclusive
⏱ 10 hrs
AI fundamentals + how AI helps in data science projects + prompt engineering for data tasks + Gemini API for AI insight generation + Copilot as a daily coding tool. Three ML paradigms and DFS/BFS algorithms. Concept-first, practical-second. No heavy math.
AI fundamentals + how AI helps in data science projects + prompt engineering for data tasks + Gemini API for AI insight generation + Copilot as a daily coding tool. Three ML paradigms and DFS/BFS algorithms. Concept-first, practical-second. No heavy math.
AI1
AI Fundamentals — Simply Explained
What is AI, ML, Deep Learning, LLMs — the hierarchy in plain English. How a language model works. Why it gives different answers each time (temperature). Difference between discriminative models (classification) and generative models (LLMs). What training data is and why it matters for data scientists working alongside AI tools.
AI Fundamentals
AI2
How AI Helps in Data Science IT Service Projects
Real scenarios: (1) Auto-generate plain-English insight summaries under a chart for a non-technical client. (2) Allow stakeholders to query a dataset in natural language. (3) Generate automated monthly report narratives from structured data. (4) Summarize customer feedback for a client dashboard. Students map each scenario to code they can build and deliver on the job.
AI in DS projectsReal scenarios
AI3
Prompt Engineering for Data Tasks
System prompt vs user prompt. Role prompting for data analysis: "You are a business analyst. Summarize this sales data in 3 bullet points for a non-technical executive." Structured output: "Always respond in JSON." Context injection: send a DataFrame summary, ask AI to interpret it. Few-shot: give example data → insight pairs so the model learns the style. Students build 5 working prompts for data science use cases.
Prompt EngineeringData promptsContext injection
AI4
Gemini API — AI Insights in Python + Streamlit
API key from Google AI Studio (free, no credit card). Call from Python. Pass a DataFrame summary as context — receive a plain-English insight paragraph. Integrate into the Streamlit app: chart renders → AI-generated insight auto-appears below it. Result: a data app that explains itself to business users.
Free tier: 15 req/min, 1M tokens/day. ₹0. No credit card needed.
Gemini APIAI insight generation
AI5
Copilot + ChatGPT + Claude — Daily DS Workflow
GitHub Copilot: write a comment describing the transformation → Tab → review before using. Best for: Pandas cleaning chains, chart boilerplate, scikit-learn pipelines. ChatGPT debug template: "Error: [X]. Code: [paste]. Expected: [Z]. Fix?" Claude for notebook review: "Check my statistical interpretation — is this conclusion valid?"
Copilot free with GitHub Student Pack. ChatGPT + Claude free tiers sufficient for all students.
GitHub CopilotChatGPTClaude
ML — Three Learning Paradigms
ML1
Supervised Learning — Deeper Theory
Bias-variance trade-off. Ensemble methods: Bagging vs Boosting. XGBoost hands-on — when it beats Random Forest. SHAP values for model explainability.
Key algorithms
- Linear Regression — predict a number
- Logistic Regression — yes/no output
- Decision Tree — if/else at scale
- XGBoost — ensemble boosting
IT service use cases
- Customer churn prediction for CRM
- Support ticket priority classification
- Price prediction for marketplace
- Lead scoring for marketing tools
ML2
Unsupervised Learning
Model finds patterns in data with no labels. k-Means: centroid logic, elbow method. DBSCAN concept. PCA applied to real high-dimensional dataset. Autoencoders for anomaly detection concept.
Key concepts
- k-Means — group by distance
- DBSCAN — density-based clustering
- PCA — dimensionality reduction
- Anomaly detection concept
IT service use cases
- Customer segmentation for marketing
- Product grouping for e-commerce
- Unusual login pattern detection
- Similar ticket auto-grouping
ML3
Reinforcement Learning
Agent learns by trial and error — takes actions, gets rewards or penalties. Taught conceptually only. Students understand why it powers recommendation engines and dynamic pricing — common interview awareness question.
Key concepts
- Agent, environment, action, reward
- Explore vs exploit trade-off
- Reward shaping
- Real-world: recommendation systems
IT service relevance
- Powers "recommended services" features
- Dynamic pricing engines
- Chatbot self-improvement basis
- Interview awareness — very common
AI Search Algorithms — DFS & BFS
ALG1
DFS — Depth First Search
Explore as deep as possible before backtracking. Stack-based traversal. Recursive DFS in Python. Decision tree paths, knowledge graph traversal, nested JSON crawling.
What students learn
- Stack-based traversal logic
- Recursive DFS in Python
- Tree traversal orders
- Time complexity: O(V+E)
DS + AI use cases
- Decision tree traversal in ML
- Nested data structure crawling
- Feature dependency graph
- Pathfinding in graph data
ALG2
BFS — Breadth First Search
Explore all neighbors at current depth before going deeper. Queue-based. Finds shortest path. Social graph analysis, level-order tree traversal, network analysis on customer relationship data.
What students learn
- Queue-based traversal logic
- BFS implementation in Python
- Level-order tree traversal
- Shortest path concept
DS + AI use cases
- Social graph analysis
- Network analysis on customer data
- BFS in AI state-space search
- Web crawler page discovery
Internship
I1
Live Data Brief — Ether Services or Partner Company
Real client engagement: analyse a client's sales or booking data, build a monthly reporting dashboard, train a prediction model for their use case, or add AI insight summaries to an existing report. Real GitHub commits on a live project. Supervised by Ether Services team. Signed internship certificate + a specific story to tell in interviews.
▸ Output: Signed certificate + GitHub activity on real project + client story for interview.
Internship
Job Descriptions — 8hr Graduates
These roles require Deep Learning, NLP, transformers, or AI awareness — all covered in the 8hr Advanced Track. Click to expand.
Data Scientist — AI/ML Focus
AI-first companies · Chennai / Bangalore / Remote · ₹4–6 LPA
Role
End-to-end ML + DL projects. Model training, evaluation, deployment. NLP, tabular, and image data. Deploy models as APIs or web apps for clients.
Required Skills
PythonTensorFlow/Kerasscikit-learnNLPSQLStatisticsDeployed project
Interview
Share live project URL. Explain model architecture, evaluation metric choice, and one improvement. Code a feature in Python live.
Search on Glassdoor →NLP Engineer / Conversational AI Developer
AI product companies · Pan India / Remote · ₹4–7 LPA
Role
Build NLP pipelines for chatbots, ticket classification, document summarization. HuggingFace models, transformers, LLM APIs. Fine-tune pre-trained models on custom data.
Required Skills
PythonHuggingFaceTransformersNLPLLM APIsPrompt engineering
Search similar on Indeed →ML Engineer — Deployment + MLOps
IT product companies · Chennai / Remote · ₹4.5–6 LPA
Role
Build, track, and deploy ML models. Maintain experiment logs with MLflow. Deploy as Flask APIs or Streamlit apps. Monitor model performance in production.
Required Skills
Pythonscikit-learnTensorFlowMLflowFlaskStreamlitGit
Search on Naukri →AI-Integrated Data Scientist — IT Services
Mid-size IT services · Pan India / Remote · ₹4–5.5 LPA
Role
Build AI-powered data features for client IT projects. Integrate LLM APIs for insight generation, add AI summaries to dashboards, build question-answering on structured data.
Required Skills
PythonPandasGemini/OpenAI APIPrompt engineeringStreamlitSQL
Search similar on Internshala →Junior Deep Learning / Computer Vision Engineer
AI startups + tech companies · Bangalore / Chennai · ₹5–8 LPA
Role
Build CNN-based image classification and object detection models. Transfer learning with pre-trained models. Model training, evaluation, and deployment.
Required Skills
PythonTensorFlow/KerasCNNTransfer learningNumPyGit
Search on Glassdoor →Program Fee
₹59,899
8hr/day · 4 months · 768 hrs · Internship + Placement support included
8hr Program Summary
4 Projects
Mini models + DataSpark + DataSpark Pro + Internship
Deep Learning
TF/Keras + CNN + LSTM + Transformers + MLflow
Live URL
Deployed Streamlit Pro app
Internship
Certificate + placement support
Both Programs
The DataSpark
Project
Project
Full data science pipeline on the Zuvio marketplace dataset — bookings, providers, customers, revenue, ratings, reviews. Interviewers immediately understand the business context, and every model output has a story attached.
2hr Core Track · 2 Projects
DataSpark Core
P1
Mini ML Models
Built inside ML modules. Cancellation predictor + price predictor trained and evaluated. Pushed to GitHub.
P2
DataSpark — Capstone
SQL → Pandas → EDA → Classification + Regression + NLP → Streamlit app → insight report. Live public URL.
8hr Advanced Track · 4 Projects
DataSpark Pro
P1
Mini ML Models
Same as Core Track P1 — built inside ML modules.
P2
DataSpark Core
Same as Core Track P2 — completed within first 3 months.
P3
DataSpark Pro — Flagship
XGBoost + Deep Learning + HuggingFace NLP + Time Series + MLflow + Gemini AI insights + Streamlit Pro. Live deployed.
P4
Internship Project
Real client data brief under Ether Services. Certificate + GitHub commits.
DataSpark Core — 2hr Capstone
Python + Pandas + scikit-learn + Streamlit + GitHub
DataSpark — Zuvio Marketplace DS Pipeline
2hr capstone · manual end-to-end data science
SQL extraction: Pull bookings, providers, customers, reviews from MySQL. Joins across 4 tables. 10 business questions answered in SQL first.
Pandas cleaning + EDA: Nulls, duplicates, datetime features, groupby summaries, correlation heatmap, outlier detection. Documented notebook with business interpretation.
Classification model: Booking cancellation predictor — Random Forest. Feature importance. Precision/Recall/F1. Plain-English explanation for a non-technical client.
Regression model: Booking price predictor — Linear Regression vs Ridge. RMSE, R² comparison.
NLP sentiment analysis: Classify reviews as positive/negative using TF-IDF + Logistic Regression.
Streamlit app: Upload dataset → EDA → cancellation risk → price estimate → review sentiment. Live on Streamlit Community Cloud.
DataSpark Pro — 8hr Flagship
XGBoost + TensorFlow + HuggingFace + Prophet + MLflow + Gemini + Streamlit
DataSpark Pro — AI-Powered DS Platform
8hr flagship · advanced models · AI insights · live deployed public URL
Advanced modelling: XGBoost cancellation predictor — beats Random Forest baseline. SHAP explainability. MLflow comparison across 5 model runs.
Deep learning: Neural network classifier via Keras. CNN for image-based category classification. LSTM for booking volume forecasting — all tracked in MLflow.
Advanced NLP: HuggingFace zero-shot ticket classifier + BERT sentiment. Compared against TF-IDF baseline — students see the quality gap in one chart.
Time series forecasting: Prophet + LSTM predict next 4 weeks of bookings. Seasonality components visualized.
AI insight generation: DataFrame summary → Gemini API → plain-English insight paragraph auto-displayed under each chart. Non-technical users understand every chart without asking a data scientist.
Live Streamlit Pro app: Multi-page. File upload → EDA → predictions → AI insights → time series forecast. Deployed on Streamlit Community Cloud — interviewer opens the link live during the interview.
8hr Advanced Track · Deployment Module
Live Hosting
Guide
Guide
Every 8hr student leaves with a permanent public URL. Taught in one dedicated session — students deploy during class, not alone at home.
Streamlit App → Streamlit Community Cloud · Free Always
Connect GitHub → auto-deploys on every push
requirements.txt: pandas, streamlit, scikit-learn, tensorflow, transformers, prophet, google-generativeai
Gemini API key stored in Streamlit Secrets — never in code
Student URL: yourname-dataspark.streamlit.app
Cost: ₹0 always
MLflow + GitHub Portfolio
MLflow experiment screenshots saved to GitHub
GitHub README links: live Streamlit URL + notebook + MLflow screenshots + insight report
DagsHub (free): host MLflow tracking server for public viewing — optional but impressive
Cost: ₹0 always
Step-by-step — one class session
1
Prepare the repo
Clean folder: data/, notebooks/, app.py, requirements.txt, README with project story and screenshots. Push to GitHub.
2
Deploy on Streamlit Community Cloud
share.streamlit.io → New app → GitHub repo → app.py. Add GEMINI_API_KEY in Secrets. Live in minutes.
3
Add MLflow screenshots to GitHub
Screenshot experiment comparison table and best run metrics. Add to /mlflow-screenshots/ in repo. Mention in README.
4
Test AI features live
Upload the sample CSV. Generate AI insights. Run cancellation prediction. Fix Secrets/requirements issues in class.
5
Portfolio + LinkedIn post
Add live Streamlit link to resume and LinkedIn. Screenshot the running app. Post: "Built a live DS app with ML predictions + AI insights on Zuvio marketplace data."
Options comparison
| Platform | Best for | Free tier | Verdict |
|---|---|---|---|
| Streamlit Community Cloud | Python DS/ML apps | Always free | Use this — built for exactly this |
| DagsHub | MLflow + Git hosting | Free public repos | Use for MLflow tracking |
| HuggingFace Spaces | Streamlit alternative | Free | Good backup if Streamlit Cloud is slow |
| Render | Flask API hosting | Free (sleeps) | For Flask model API if needed |
8hr Advanced Track Exclusive
AI Tools Guide
Two categories: tools that make students analyse and code faster, and APIs to add AI insight features to their project. Total student cost: ₹0 using free tiers and GitHub Student Pack.
Coding + analysis tools
T1
GitHub Copilot
Best for: Pandas cleaning chains, Matplotlib chart code, scikit-learn pipelines, repetitive evaluation boilerplate. Write a comment describing the transformation → Tab to accept → always read before using.
Free with GitHub Student Developer Pack (education.github.com). Otherwise $10/month.
T2
ChatGPT
Debug template: "Error: [X]. Code: [paste]. Dataset shape: [Y]. Expected: [Z]. Fix?" Also: explain sklearn docs, draft README, generate realistic test data, interpret statistical output in plain English.
Free tier sufficient. Faculty uses one Plus plan for class demos.
T3
Claude
Best for: "Review my analysis notebook — any logic issues?" "Is my statistical interpretation correct?" More thorough for reasoning review. Excellent for polishing insight reports.
Free tier at claude.ai sufficient. Faculty uses one Pro plan for demos.
AI APIs for project features
A1
Google Gemini API — teach first
Free tier: 15 req/min, 1M tokens/day. API key from aistudio.google.com — no credit card. Used for: auto-generating insight paragraphs from DataFrame summaries, natural language data Q&A in the Streamlit app.
Cost: ₹0. No credit card. Every student uses from day one of the AI module.
A2
OpenAI API — optional
GPT-4o-mini: ~$0.60 per 1M tokens. ~₹50 total for student project. Teach as alternative — needed for JDs that specifically mention OpenAI.
Cost: ~₹25–50 total. Optional.
Total cost — 4 months
| Tool | Plan | Student cost | Note |
|---|---|---|---|
| GitHub Copilot | Student Pack | ₹0 | Free with student verification |
| ChatGPT | Free tier | ₹0 | Sufficient for all use cases |
| Claude | Free tier | ₹0 | Sufficient for notebook review |
| Gemini API | Free tier | ₹0 | No credit card. All AI features. |
| Streamlit Cloud | Community | ₹0 | Free app hosting, always on |
| Total | ₹0 for entire program | ||
Both Programs
Schedule &
Duration
Duration
2hr Core Track — 3 months, Mon–Fri, 120 hrs. 8hr Advanced Track — 4 months, Mon–Sat, 768 hrs. Both programs include placement support.
Program Overview
2hr/day · Core Track
3 months · Mon–Fri · 120 hrs
Python + NumPy + Pandas + Statistics + Linear Algebra intuition + SQL + ML (Regression, Classification, Clustering, NLP) + Feature Engineering + Streamlit + Git
2 projects — Mini ML models + DataSpark capstone with live Streamlit URL
No AI shortcuts — pure manual analysis and modelling
Placement support — resume prep, mock interview, company referrals
₹29,899
8hr/day · Advanced Track · Hero Product
4 months · Mon–Sat · 768 hrs
Months 1–3: Complete 2hr Core Track curriculum
Month 4 onwards: Deep Learning (TF/Keras, CNN, LSTM) + Advanced NLP (HuggingFace, Transformers) + Time Series + MLflow + AI fundamentals + ML paradigms + DFS/BFS + Prompt engineering + Gemini API + Copilot + DataSpark Pro deployment + Internship
4 projects — Mini models + DataSpark + DataSpark Pro (live URL) + Internship project
Internship certificate from Ether Services
Placement support — resume, mock interview, company referrals
₹59,899
For enquiries contact Mohan @ 9597271533 · www.appincoimbatore.com