Hello, I'm Aryan Deshpande

I work on-

I develop cloud-native data pipelines, train and serve ML Systems, and build backend & app services.

View Projects Download Resume
☁ ETL ML Data Lake API > train_model() Python Docker AWS XGBoost

About Me

I'm a Final year B.Tech Information Technology student at VIT Vellore, India, focused on Cloud & Data Engineering. I build end-to-end data systems, from ETL ingestion, storage and processing, to model training and real-time inference, with an eye on cost-efficiency, performance and maintainability. I also develop backend services and mobile apps when projects need full-stack delivery.

ML Production Cloud Architecture Data Processing

Quick Stats

8.72
CGPA
5+
Major Projects
AWS
Certified

Need the resume?

Download or view on Drive

Download Resume Drive

Skills

Core capabilities across cloud & data, ML systems, backend services and Flutter app development.

Cloud & Data

S3, Athena, Glue, Lambda, SageMaker, Medallion architecture, ETL & query optimization.

AWS VPC Architecture Parquet ETL Dashboarding

ML Systems

Model training & deployment (XGBoost, Scikit-learn, SageMaker, FastAPI), feature engineering.

Python XGBoost RDKit joblib

Backend

APIs and production services using Express.js, PostgreSQL, JWT, Docker, and REST best practices.

Express.js Postgres Docker REST

App Dev

Flutter apps with Maps integration, storage, and performant UI patterns for Android/iOS.

Flutter & Firebase BLoC Architecture

My Work

Recent Projects

SEC

Dockerized Zeek Pipeline for DNS Exfiltration Detection

Built a containarized DNS threat-detection pipeline using Zeek, FastAPI, and Random Forest ML to analyze raw PCAP traffic and detect DNS exfiltration attacks with 98% malicious recall, leveraging advanced telemetry features like Shannon entropy and subdomain analysis for protocol-aware forensic detection.

Tech: Zeek, Docker, FastAPI, Python, Random Forest, PCAP Analysis
GitHubLive Link
Cloud

Data Processing Architecture on AWS

Developed an ML pipeline (including ETL) with S3 partitioning and Parquet, improved Athena query performance by ~70%, trained XGBoost on SageMaker and deployed real-time inference via FastAPI on EC2 with Docker. Serverless ETL reduced infra costs significantly.

Tech: AWS (SageMaker, Glue, Athena, Lambda), Python, XGBoost, FastAPI
GitHubDemo
APP

Geo-PhotoVault App

Flutter app that auto-organizes photos into city-based folders using GPS metadata, with Google Maps visualization and location filters for fast retrieval.

Tech: Flutter, Dart, Google Maps API
GitHubDemo
API

Hackulus '25 Backend Portal

Built the event backend (150+ participants) with Express.js and PostgreSQL. Implemented JWT auth and admin workflows to automate tasks and maintain >99% uptime during the event.

Tech: Express.js, PostgreSQL, JWT, Redis, Sentry
GitHub Live Link
ML

Chemical Reaction Catalyst Predictor

ML pipeline using RDKit Morgan fingerprints on an SGD classifier to predict catalysts; improved recall and reduced physical experiment trials. Employed joblib for fast inference.

Tech: Python Pandas, RDKit, Scikit-learn, FastAPI
GitHubLive Link

Credentials

Certifications

AWS
AWS Certified Cloud Practitioner
CLF-C02
View
IBM
IBM Watsonx - Generative AI
CEWXAI1IN
View

My Roles

Technical and Leadership Experience

Projects Head - SIAM, VIT
Jan 2025 - Present • Vellore, India
SIAM logo

Led cross-functional technical projects: coordinated dev & design teams, conducted technical interviews, mentored juniors, and ensured code quality. Drove process improvements and managed financial operations.

Leadership Technical Hiring Project Ops & Finance
Core Committee Member (App Dev) - IEEE CS, VIT
Jan 2024 - Jan 2025 • Vellore, India
IEEE CS VIT

Worked on modular Flutter components, performance tuning, and implemented self-check mechanisms to catch regressions early. Participated in code reviews to boost app stability.

Flutter Performance Optimisation Code Review

Get In Touch

Contact Me

Want to collaborate or see a technical walkthrough? Drop a message and I'll get back within 48 hours.

Or email me directly at dshryng@gmail.com
Connect on GitHub • LinkedIn