Software Engineering Internship Training Report | Sonata

SCHOOL OF COMPUTER SCIENCE ENGINEERING

INTERNSHIP

TRAINING

4-Week Onboarding & Technical Training Programme

28 Training Days

20+ Modules Covered

6 Hackathon Team Members

Rajesh D

01FE23BCS429

Dr. Guruprasad S K

Sonata Software, Bengaluru

2025–2026

PROGRAMME OVERVIEW

Programme Overview

Days 1–2

Induction & Orientation

Company culture, AGILE values, organisational structure, IT onboarding, and team introductions.

Days 3–28

Technical Training

20+ modules spanning software engineering, databases, cloud, AI, web development, and professional skills.

Capstone

Hackathon

Cross-functional team of 7 built a working AI-powered Vendor Quote Rating Agent in a condensed sprint.

COMPANY ORIENTATION

AGILE Core Values

A

Action

Taking initiative, moving with purpose, and delivering results

G

Growth

Continuously improving personally, professionally & as an organisation

I

Integrity

Being honest, transparent and accountable in everything we do

L

Learning

Embracing curiosity and treating every challenge as an opportunity

E

Empathy

Understanding and respecting the perspectives of all stakeholders

Industry Verticals: BFSI · HLS · TMT · RMD

DOMAIN EXPERTISE

INDUSTRY VERTICALS & SERVICE AREAS

BFSI

Banking, Financial Services & Insurance

Compliance, risk management & digital transformation for financial institutions.

HLS

Healthcare & Life Sciences

Data, compliance & operational systems for healthcare & pharma organisations.

TMT

Technology, Media & Telecom

Scalable platforms & digital services for tech and media companies.

RMD

Retail, Mfg & Distribution

Efficiency and modernisation for retail and supply chain operations.

Core Service Areas

☁ Cloud

📊 Data

🤖 Artificial Intelligence

⚙ Microsoft Dynamics

🔄 Digital Transformation

TRAINING OVERVIEW

TECHNICAL TRAINING MODULES

Days 3–28

Software Eng & Agile

Git & Version Control

SQL & NoSQL DBs

Security & Infrastructure

Cloud & Distributed Sys

HTML / CSS / JavaScript

Python, Pandas & NumPy

REST APIs & Flask

Testing & TDD

AI/ML & Prompt Eng

Agentic AI & Agent Building

Architecture & Pro Skills

M O D U L E D E E P D I V E

Software Engineering & Agile Methodology

SDLC Phases

Requirements Gathering

System Design

Implementation

Testing & QA

Deployment

Maintenance

Scrum Ceremonies

Sprint Planning

Define sprint goals & break down tasks for the team

Daily Stand-up

15-min sync: what I did, what I'll do, any blockers

Sprint Review

Demo completed work to stakeholders for feedback

Retrospective

Reflect as a team and improve processes going forward

Key Takeaway:

Clean code, peer reviews, and meaningful commit messages are as important as the feature itself.

MODULE DEEP DIVE

Databases: SQL & NoSQL

SQL — Relational Databases

Joins (INNER, LEFT, RIGHT, FULL)

Subqueries & Common Table Expressions

Indexing & query optimisation

Normalisation — 1NF through 3NF

Stored procedures & views

ACID properties & transactions

NoSQL — Non-Relational Databases

Document stores — MongoDB

Key-value & column-family stores

Graph databases & use cases

CAP Theorem — Consistency, Availability, Partition

When to choose SQL vs NoSQL

Schema-less design patterns

M O D U L E D E E P D I V E

Security, Cloud & Distributed Systems

OWASP Top Vulnerabilities

Cloud Service Models

Injection Attacks (SQL, XSS)

Broken Authentication

Sensitive Data Exposure

Security Misconfiguration

Insecure Deserialization

IaaS — Infrastructure as a Service

VMs, storage, networking — e.g. AWS EC2

PaaS — Platform as a Service

Managed runtimes — e.g. Azure App Service

SaaS — Software as a Service

Ready-to-use apps — e.g. Microsoft 365

MODULE DEEP DIVE

AI / ML, Prompt Engineering & Agentic AI

Machine Learning Lifecycle

Data Collection

Data Cleaning

Feature Engineering

Model Training

Evaluation

Deployment

Prompt Engineering Techniques

Zero-Shot Prompting

No examples — direct instruction to the model

Few-Shot Learning

Providing examples to guide model output

Chain-of-Thought (CoT)

Asking the model to reason step-by-step

System Prompt Design

Setting role, tone & constraints upfront

Agentic AI Concepts

ReAct Pattern — Reasoning + Acting in loops

Tool use & function calling

Multi-agent orchestration

LangChain framework basics

Memory, planning & reflection loops

M O D U L E D E E P D I V E

Web Development · Python · REST APIs

Front-End Web Dev

Semantic HTML5 structure

CSS Flexbox & Grid layouts

Responsive design principles

JavaScript ES6+ features

DOM manipulation & events

Interactive UI with vanilla JS

Python & Data

Python syntax & data types

Pandas — load, clean, transform

NumPy — numerical operations

Matplotlib visualisations

Exploratory Data Analysis

Scripting & automation

REST APIs & Flask

HTTP methods — GET, POST, PUT, DELETE

CRUD endpoint design

Request validation & error handling

Authentication — API keys, JWT

API versioning strategies

Testing with Postman

MODULE DEEP DIVE

Testing & Test-Driven Development

Red → Green → Refactor Cycle

Testing Pyramid

Write a failing test for new functionality

Write minimum code to make the test pass

Clean & improve code without breaking tests

E2E Tests

Simulate real user flows

Integration Tests

Test component interactions

Unit Tests

Fast, isolated, most numerous

Write the test FIRST — always

Code coverage measures quality breadth

Unit > Integration > E2E in quantity

H A C K A T H O N C A P S T O N E

AI Agent for Vendor Quote Rating

AI-powered procurement intelligence — an ML system that automates vendor quote evaluation and scoring.

01

Quote Upload

Procurement users upload vendor quote documents for processing.

02

ML Scoring

Random Forest Regressor predicts a rating score; Classifier categorises: Recommended, Acceptable, or Rejected.

03

Risk Analysis

Risk Engine evaluates risk factors and the Recommendation Engine generates procurement guidance.

04

LLM Justification

Gemini API & LLaMA3 generate human-readable justifications via an interactive AI Chat Interface.

React

Python

Flask

Random Forest

Gemini API

LLaMA3

Pandas

NumPy

H A C K A T H O N

System Architecture — How It Works

User Input

Procurement user uploads vendor quotes via the React frontend dashboard.

Extraction & Normalisation

Data extracted and normalised from different vendor formats for fair comparison.

ML Scoring Engine

Random Forest Regressor scores each quote; Classifier assigns: Recommended / Acceptable / Rejected.

Risk Engine

Evaluates risk factors; Recommendation Engine generates procurement guidance.

LLM Justification

Gemini API & LLaMA3 generate natural language justifications for each rating.

Dashboard Output

Results, KPIs, risk scores, and AI chat interface displayed on Analysis Dashboard.

Client Layer (React) → API Layer (Flask) → Processing Engine → AI/ML Layer → Data Layer

R E F L E C T I O N S

Key Learnings from 4 Weeks

Agile in Practice

Scrum ceremonies & sprint delivery gave real-world context to SDLC theory and reinforced iterative thinking.

Full-Stack Depth

From SQL to NoSQL, REST APIs to front-end — a holistic view of how modern software is built end to end.

AI at the Frontier

Agentic AI, LangChain & the ReAct pattern offered a glimpse into systems that reason, plan and act autonomously.

Hands-On Learning

The hackathon and coding exercises bridged the gap between conceptual knowledge and deployable applications.

Security-First Mindset

OWASP Top 10 and PowerShell scripting reinforced that security must be embedded throughout development.

Professional Skills

Communication, stakeholder management & code review discipline are equally critical career skills as technical depth.

C E R T I F I C A T I O N S

SCALE Certifications Completed

Courses completed via SCALE — Sonata Career Academy for Learning Excellence

Compliance

POSH

Prevention of Sexual Harassment at the Workplace 2025-26

11 Mar 2026 · 1 hr

Artificial Intelligence

AI & Machine Learning

Artificial Intelligence and Machine Learning

25 Mar 2026 · 1 hr 36 m

Python

Python Development

Leveraging Functions with Lambdas, Generators, Closures & Decorators

27 Mar 2026 · 1 hr 31 m

Prompt Engineering

ChatGPT Prompts

Finetuning Your ChatGPT Prompts

27 Mar 2026 · 1 hr 19 m

Generative AI

Generative AI

An Introduction to Generative AI

27 Mar 2026 · 1 hr 40 m

Total Learning Hours: ~7 hrs

All certifications issued by Sonata Career Academy for Learning Excellence

Thank You

Four weeks of intensive learning — from Agile foundations to Agentic AI — have built a strong foundation for a career in modern software engineering.

Software Eng

Git

Databases

Security

Cloud

Web Dev

Python

REST APIs

Testing

AI/ML

Agentic AI

Architecture

Rajesh D  |  01FE23BCS429  |  KLE Technological University, Hubballi

Internship Training Report  ·  4-Week Programme  ·  Sonata Software, Bengaluru  ·  2025–2026