Final Year · Computer Science · Cloud & AI

Puru Raj
Dhama

Building intelligent systems in the cloud
from neural networks to distributed infrastructure.

Scroll to explore

About

Puru Raj Dhama portrait
Puru Raj Dhama

I'm Puru Raj Dhama, a final-year Computer Science student at Bennett University (Times Group), Greater Noida, specialising in Cloud Computing and Artificial Intelligence. I design and build systems that scale — from training ML models to architecting cloud-native applications on AWS, Azure and GCP.

Currently exploring the intersection of large language models and cloud infrastructure, building tools that make AI more accessible, reliable, and production-ready.

7+
Projects
1+
Internships
LC Solved
Cloud ComputingMachine Learning AWS / GCPPython JavaTensorFlow KubernetesDocker FastAPIReact PostgreSQLTerraform CI/CDLLMs / RAG

Experience

Where I've worked.

Work Experience
Jun 2025 - Jul 2025 · 2 mos
Summer Intern
Crest Data · Ahmedabad, Gujarat, India · On-site
Built custom MCP servers for vulnerability detection and prioritization. Integrated AI agents, including Windsurf and Copilot, to test real-time vulnerability remediation workflows. Designed Cascade rules and workflows for LLM-guided code fixes and introduced context engineering to evaluate prompt strategies and agent performance.
MCPContext EngineeringPrompt EngineeringAI Agents

Education

Academic journey.

2020
Class 10
Delhi Public School · CBSE Board
Grade: 89%.
2022
Class 12
Delhi Public School · CBSE Board
Grade: 87.2%.
2022 - 2026
B.Tech, Computer Science & Engineering
Bennett University (Times Group), Greater Noida
Specialisation in Cloud Computing and AI. CGPA: 7.5/10. Coursework includes Distributed Systems, Deep Learning, Cloud Architecture, and Algorithm Design.
CGPA 7.5Cloud Spec.AI Spec.

Certifications

Professional learning.

Apr 2024
AWS Cloud Practitioner Essentials
Amazon Web Services (AWS) · Credential ID HR78Z2TKVKWY
Foundational certification covering core AWS services, cloud concepts, pricing, security, and architecture fundamentals.
AWSCloud Fundamentals
Oct 2024
Google Cloud Computing Foundation with Kubernetes
Google Cloud Skills Boost · Credential ID 11810854
Covered Google Cloud fundamentals with Kubernetes-based application deployment and orchestration.
GCPKubernetes
Jun 2025
MCP: Build Rich-Context AI Apps with Anthropic
DeepLearning.AI
Learned how to build rich-context AI applications using the Model Context Protocol and Anthropic's ecosystem.
MCPAnthropicAI Apps

Projects

Things I've built.

001
Companion — Gamified accountability platform
A gamified goal-tracking platform built around social accountability circles. Users create or join circles with a shared goal, add personal daily tasks, check in every day, and compete on a live leaderboard. Designed for small groups working toward a shared commitment.
Next.jsSpring BootPostgreSQLTailwindJWT
002
AZURE-CLOUD — Azure labs & proof-of-work repository
A hands-on collection of Microsoft Azure labs, notes, and implementations demonstrating core cloud concepts through practical, real-world scenarios. Covers compute, storage, networking, identity/access, databases, and monitoring/security — built as a "proof of work" documenting applied cloud skills.
Microsoft AzureAzure AD / RBACAzure Networking & Storage
003
Yapo — Serverless blog app on AWS
A modern serverless blog application with user sign-up/login and the ability to create, view, and comment on posts via a protected authentication flow. Built end-to-end on AWS managed services, with AI features including Text-to-Speech (Polly) and Translation (Translate) baked in.
AWS LambdaAPI GatewayDynamoDB + S3AWS Cognito
004
BottlelineClassification — Bottle fill-level image classification
A deep-learning computer vision project that classifies bottle images as Underfilled, Overfilled, or Normal to support automated industrial inspection. Trained a custom CNN and MobileNetV2 transfer-learning model, reaching ~98% accuracy vs a ~71% baseline. Published in IEEE.
PythonTensorFlow / KerasMobileNetV2OpenCVScikit-learn
005
malware-detection — Static malware classification
A machine-learning system that classifies malware families using static-analysis features extracted from PE files, DLL imports, and API calls. Implements an XGBoost pipeline with feature selection and cross-validation achieving ~93% accuracy, plus a Streamlit demo for interactive predictions.
PythonXGBoostScikit-learnStreamlit

Achievements

Recognition & milestones.

2025
"Automated Detection of Bottle Fill Levels Using CNN and MobileNetV2" published in IEEE.
2025
Winner - Mined, Crest Data @ Nirma University
Hackathon win for a solution focused on practical problem-solving and technical execution.
2025
Finalist - Datathon by Analytika @ NMIMS
Reached the finalist stage in a competitive datathon with a strong analytical and implementation-focused submission.
2025
"Serverless Deployment of an Ensemble ML Model for Diabetes Prediction" — published at AECE 2025 (IEEE).

LeetCode

Consistent practice.

PR
Raj_0106
Loading profile…
Total Solved
Easy
Medium
Hard
Global Rank
↗ Profile
Submission activity · Last 52 weeks
Less
More

Contact

Let's build something
remarkable.

Open to full-time roles, internships in Cloud Engineering, ML Engineering, or Software Engineering. Also available for research collaborations and open-source projects.