Grow Data SKills | Most Affordable Courses | Become Top Data Professional

Light
Dark

AI Engineering Bootcamp

By Shubhankit In Grow Data Skills

Live Classes (In Progress)
English
Completion Certificate

Demo Class Video

1 Year

Validity (From the date of Enrollment)

70 Hours

Duration

28

Sessions

13

Projects

INR 7000

Admission Open

Buy Now

This course includes

Content Duration70 Hours
Total Live Session28
Total Industry Project13
Doubt solving in live classes and offline doubt support on private discord group
Quality Assignments & quizzes after each module
Interview preparation guide
Resume & Linkedin profile making
Dedicated placement assistance
Certificate of completion

Tech stack you'll learn

Course Content

✅ Class 1 - ML Workflow & Regression Systems

AI System Design: Data → Features → Model → API → Monitoring
DVC & MLflow integration for reproducibility
Bias–variance, under/overfitting, evaluation metrics

✅ Class 2 (Build): Project 1 – Student Score Prediction System

Class 2 (Build): Project 1 – Student Score Prediction System
Track metrics with MLflow experiments
Visualize error distributions and regression line

✅ Class 3 - Supervised ML & Deployment-Ready Classifiers

Classification models: Logistic, Random Forest, XGBoost
Model interpretability with SHAP/LIME
Architecture: train → validate → serve

✅ Class 4 (Build): Project 2 – Customer Churn Classifier API

Train Random Forest model
Integrate MLflow model registry
Deploy via FastAPI endpoint

✅ Class 5 - Unsupervised ML & Feature Engineering Pipelines

KMeans, DBSCAN, PCA, UMAP
Handling missing data, encoding, scaling
Feature Store concepts (offline + online)

✅ Class 6 (Build): Project 3 – Customer Segmentation System

Create clustering pipeline with PCA visualization
Automate preprocessing workflow
Deploy via Streamlit dashboard

✅ Class 7 - Classical NLP & Text Pipelines

Text cleaning, tokenization, stopword removal
TF-IDF, Bag of Words, Word2Vec, SBERT embeddings
Sentence similarity using cosine distance

✅ Class 8 (Build): Project 4 – Spam/Ham Email Classifier

Train TF-IDF + Logistic model
Build FastAPI endpoint for classification
Visualize most significant tokens per class

✅ Class 9 - Sequential Models & Deep Text Understanding

RNNs, LSTMs, GRUs, Attention mechanism
Vanishing gradient issue and embedding layer tuning
LSTM vs Transformer performance tradeoffs

✅ Class 10 (Build): Project 5 – Sentiment Analysis with LSTM

Train LSTM on IMDB dataset
Add attention visualization layer
Serve predictions via API

✅ Class 11 - CNNs & Model Optimization

CNN architecture, filters, pooling
Transfer learning (VGG, ResNet, EfficientNet)
Quantization, pruning, and ONNX conversion

✅ Class 12 (Build): Project 6 – Image Classifier System

Fine-tune ResNet model
Export to ONNX format
Deploy inference endpoint + visualize GradCAMs

✅ Class 13 - Transformer Architecture & Seq2Seq Models

Self-attention, encoder-decoder structure
Transformer math and position encodings
BLEU evaluation metric

✅ Class 14 (Build): Project 7 – Machine Translation System

Build Seq2Seq + Attention model for EN→FR translation
Compute BLEU & perplexity scores
Deploy inference API

✅ Class 15 - Model Serving, Versioning & Tracking

MLOps lifecycle and architecture
MLflow registry + FastAPI deployment pattern
Dockerization and environment reproducibility

✅ Class 16 (Build): Project 8 – LSTM Sentiment API with MLflow Tracking

Containerize LSTM model
Track runs, metrics, and artifacts in MLflow
Deploy API using Docker Compose

✅ Class 17 - CI/CD, Scaling, and Monitoring

Jenkins & GitHub Actions CI/CD
Kubernetes deployment (Pods, Services, Ingress, HPA)
Monitoring with Prometheus & Grafana

✅ Class 18 (Build): Project 9 – CNN Model on K8s with CI/CD

Automate Docker build + push
Trigger Jenkins job → K8s rollout
Dashboard live metrics via Grafana

✅ Class 19 - Foundation Models & Prompt Engineering

GPT, BERT, LLaMA, Granite (IBM WatsonX) internals
Transformer deep dive (multi-head attention)
Prompt engineering (few-shot, CoT, self-ask, RePrompting)
Hallucination control techniques

✅ Class 20 (Build): Project 10 – Fine-Tune GPT-2 for Text Generation

Fine-tune GPT-2 using LoRA/QLoRA
Compare outputs with baseline GPT-2
Add structured prompt evaluation UI

✅ Class 21 - Multimodal GenAI & Vision-Language Models

CLIP, BLIP, LLaVA, Gemini, Granite VLM
Image-text alignment, embeddings, and fusion
Visual-Language system design

✅ Class 22 (Build): Project 11 – Multimodal Caption Generator

Combine CLIP encoder + GPT decoder
Generate captions for images
Serve via Streamlit + API backend

✅ Class 23 - RAG Systems with Docling & PyMuPDF

Traditional RAG vs Agentic RAG vs CAG (Context-Aware Generation)
Chunking, embeddings, and retrieval flow design
Tools: Docling, IBM Granite, PyMuPDF, FAISS/Chroma/Pinecone
Document parsing pipelines with visual and tabular data extraction

✅ Class 24 (Build): Project 12 – RAG Chatbot over PDFs (with Docling & Granite)

Extract text, tables, and layout data using PyMuPDF and Docling
Process embeddings with IBM Granite models
Store vectors in Chroma/Pinecone
Build chatbot with contextual response grounding and source attribution

✅ Class 25 - Agentic RAG & Orchestration

Agentic RAG & CAG architecture
Re-ranking (BM25, ColBERT, hybrid retrieval)
Evaluating hallucination and grounding quality

✅ Class 26 (Build): Project 13 – Multi-Agent Research Assistant

Agents: Retriever, Summarizer, Synthesizer
Use Docling + Granite for contextual understanding
Store agent memory in vector database
Evaluate context precision and hallucination rate

✅ Class 27 - AI Virtual Manager System (Capstone)

MCP Server architecture (Memory, Code, Planning loop)
Multi-agent orchestration with LangGraph / CrewAI
Observability (LangSmith, LangFuse, AgentOps)
AI System Design: cost, latency, scalability trade-offs

✅ Class 28 (Build): Capstone Project – AI Virtual Manager System

Data Pipeline → ML Model → CI/CD → RAG (Docling + Granite + PyMuPDF) → Agentic Orchestration (MCP)
Planner Agent → Analyst Agent → Chat Agent
Observability dashboards using LangSmith + Grafana
End-to-end deployment on Kubernetes

Attention Seeking Resume Preparation and Interview Strategies
Strategies To Crack Tech Interviews
Linkedin Profile Making
How To Expand Your Professional Network On Linkedin
How To Use Various Job Portals
How To Approach For Referrals

Course Schedule

Course Start Date:

29-Nov-2025 (In Progress)

Mode of the Course:

Live

Course Duration:

70 Hours

Total Sessions:

Total Projects:

Validity:

1 Year (Starting From The Date Of Enrollment)

Class Timing:

Saturday & Sunday [6:00 PM - 8:00 PM Live Teaching, 8:00 PM to 8:45 PM Live Doubt Session] (IST)

Class Recording Provided:

Yes

Prerequisite:

Basic Python , Basic SQL

⚠️ Important Notice :

The video may not work on Linux due to DRM restrictions. It is only accessible on Chrome when using Windows or macOS.

Workaround: To access the video on Linux, you can create a Windows virtual machine (VM) and watch the video through the VM. Alternatively, you can use our Android or iOS application to view the video on your mobile device.

Instructor

Shubhankit Sirvaiya

AI Engineer @ Fidelity International Ex - World Wide Technology (WWT) & Accenture

Shubhankit Sirvaiya is an AI Engineer at Fidelity International, with prior experience as an AI Engineer at World Wide Technology (WWT) and as a Data Scientist at Accenture Growth & Strategy. He has 5+ years of hands-on experience across Data Science, Machine Learning, MLOps, and AI systems, working on real-world, production-grade problems. Alongside industry work, he has been teaching and mentoring for the last 4+ years, helping 1000+ students and professionals transition into Data Science and AI roles. He also runs an active LinkedIn community of 58k+ followers, where he shares practical insights around AI engineering, careers, and real industry expectations. His teaching style focuses on story-based learning and connecting the dots — breaking down complex AI concepts into clear mental models so learners don’t feel lost in the vast and often overwhelming AI syllabus. His goal is not just to teach tools, but to help learners understand why things work and how they fit together in real systems.