Applied AICollège Unica
3.3SpecializationSession 3 · Deep Learning & Applied Machine Learning

Data transformation and manipulation

20classes60htotalTH2 · PR2 · PNW2weighting per class

Class-by-class breakdown

20 classes

TH = theory · PR = practical · PNW = personal work. These are ministry weighting codes (not hours) used to split each class's minutes. Class length = course hours ÷ class count; the three time-boxes below always sum to that length.

0%

Quiz progress

0 of 0 classes attempted

First 2 classes free with a referral 19 more with enrollment180 min per classTheory 60mPractical 60mPersonal Work 60m
1

Why Transformation Is the Hidden Engine of Every Good Agent

FreeTH2PR2PNW2

An AI Integration Manager knows raw data rarely feeds an agent directly — transformation is what turns messy sources into retrieval-ready knowledge.

Theory60m
  • The transformation layer between collection and retrieval/generation
  • How bad transformations cause hallucinations and missed answers
  • Map one business dataset to the transformations an agent would need
Practical60m
  • Inspect a raw dataset and list 5 transformations needed before RAG use
  • Sketch the transform stage in a collect→transform→embed→retrieve pipeline
Personal Work60m
  • Write a one-page transformation plan for a sample agent knowledge base
  • Micro-task: submit your plan
2

Cleaning Text for GenAI — Stripping Noise from Documents

LockedTH2PR2PNW2

Locked class preview

A Data/Business Analyst who cleans text well directly improves RAG answer quality — noise in means garbage answers out

Enroll to unlock every class in this course — the full agenda, the AI tutor, and the quiz.

Talk to admissions
3

Splitting and Chunking — Designing Retrieval Units for RAG

LockedTH2PR2PNW2

Locked class preview

A Project Manager who understands chunking knows it's the single biggest lever on RAG quality — chunk size changes what the agent can find

Enroll to unlock every class in this course — the full agenda, the AI tutor, and the quiz.

Talk to admissions
4

Metadata Enrichment — Tagging Chunks for Smarter Retrieval

LockedTH2PR2PNW2

Locked class preview

An AI Integration Manager who tags chunks with metadata enables filtered retrieval — the agent can search only 'policy' or '2025' docs

Enroll to unlock every class in this course — the full agenda, the AI tutor, and the quiz.

Talk to admissions
5

Deduplication and Near-Duplicate Detection

LockedTH2PR2PNW2

Locked class preview

A Data/Business Analyst who dedupes prevents the agent from retrieving the same passage five times — duplicates skew answers and waste context

Enroll to unlock every class in this course — the full agenda, the AI tutor, and the quiz.

Talk to admissions
6

Structuring Unstructured Data — Tables, QA Pairs and Schemas

LockedTH2PR2PNW2

Locked class preview

A Project Manager who turns unstructured docs into structured QA pairs gives an agent cleaner, more retrievable knowledge

Enroll to unlock every class in this course — the full agenda, the AI tutor, and the quiz.

Talk to admissions
7

Feature Engineering for ML Models That Support Agents

LockedTH2PR2PNW2

Locked class preview

A Data/Business Analyst who engineers features well improves the classical ML models that often sit alongside GenAI in a business solution

Enroll to unlock every class in this course — the full agenda, the AI tutor, and the quiz.

Talk to admissions
8

Building Transformation Pipelines — Repeatable, Not One-Off

LockedTH2PR2PNW2

Locked class preview

An AI Integration Manager who builds repeatable pipelines keeps the agent's knowledge base consistent as data refreshes

Enroll to unlock every class in this course — the full agenda, the AI tutor, and the quiz.

Talk to admissions
9

Embedding Pipelines — Turning Chunks into Vectors at Scale

LockedTH2PR2PNW2

Locked class preview

A Data/Business Analyst who runs an embedding pipeline can populate a vector store with hundreds of chunks automatically

Enroll to unlock every class in this course — the full agenda, the AI tutor, and the quiz.

Talk to admissions
10

Transforming Tool Outputs — Normalizing Agent API Results

LockedTH2PR2PNW2

Locked class preview

A Project Manager who normalizes tool outputs lets an agent combine results from different APIs into one coherent answer

Enroll to unlock every class in this course — the full agenda, the AI tutor, and the quiz.

Talk to admissions
11

Data Joining and Merging — Combining Sources for Richer Retrieval

LockedTH2PR2PNW2

Locked class preview

A Data/Business Analyst who merges a product catalog with support docs gives an agent richer context to answer from

Enroll to unlock every class in this course — the full agenda, the AI tutor, and the quiz.

Talk to admissions
12

Aggregation and Summarization — Compressing Data for Context

LockedTH2PR2PNW2

Locked class preview

An AI Integration Manager who summarizes large datasets into compact stats fits them into an agent's limited context window

Enroll to unlock every class in this course — the full agenda, the AI tutor, and the quiz.

Talk to admissions
13

Handling PII and Sensitive Data During Transformation

LockedTH2PR2PNW2

Locked class preview

A Data/Business Analyst who redacts PII during transformation keeps the agent from retrieving and exposing private data

Enroll to unlock every class in this course — the full agenda, the AI tutor, and the quiz.

Talk to admissions
14

Validation and Quality Checks in the Pipeline

LockedTH2PR2PNW2

Locked class preview

A Project Manager who adds validation gates catches transformation errors before they poison the agent's knowledge base

Enroll to unlock every class in this course — the full agenda, the AI tutor, and the quiz.

Talk to admissions
15

Versioning Transformed Data — Reproducible Agent Knowledge

LockedTH2PR2PNW2

Locked class preview

An AI Integration Manager who versions transformed data can explain exactly what the agent knew at any point in time

Enroll to unlock every class in this course — the full agenda, the AI tutor, and the quiz.

Talk to admissions
16

Monitoring Pipeline Health — Drift, Failures and Freshness

LockedTH2PR2PNW2

Locked class preview

A Data/Business Analyst who monitors the pipeline catches data drift before the agent starts giving stale or broken answers

Enroll to unlock every class in this course — the full agenda, the AI tutor, and the quiz.

Talk to admissions
17

Mini-Project — Build a RAG-Ready Transformation Pipeline

LockedTH2PR2PNW2

Locked class preview

This ties transformation together — a Data/Business Analyst who builds a clean, validated, embeddable pipeline proves the agent's knowledge is trustworthy

Enroll to unlock every class in this course — the full agenda, the AI tutor, and the quiz.

Talk to admissions
18

Evaluating Transformation Impact on Agent Answers

LockedTH2PR2PNW2

Locked class preview

An AI Integration Manager who measures how transformations change agent answers can prove the work was worth it

Enroll to unlock every class in this course — the full agenda, the AI tutor, and the quiz.

Talk to admissions
19

Streaming and Real-Time Transformation for Live Agents

LockedTH2PR2PNW2

Locked class preview

An AI Integration Manager who transforms data in real time lets an agent act on fresh events — streaming pipelines power live business agents

Enroll to unlock every class in this course — the full agenda, the AI tutor, and the quiz.

Talk to admissions
20

Present and Reflect — Your Transformation Pipeline

LockedTH2PR2PNW2

Locked class preview

A Data/Business Analyst presenting a transformation pipeline proves the agent's answers rest on clean, well-structured data

Enroll to unlock every class in this course — the full agenda, the AI tutor, and the quiz.

Talk to admissions

Unlock all 20 classes

Get the full agenda, AI tutor, and quiz for every class in 3.3 · Data transformation and manipulation.

Apply via admissions

Online checkout not yet live — contact admissions to enroll.

Previous course

3.2 · Programming with R for data analysis and visualization

Next course

3.4 · Supervised learning algorithms

Back to Session 3