Aditya Chaturvedi
Building scalable AI systems, intelligent backend infrastructure, and research-driven applications that push the boundary of what software can do.
Who I Am
I'm a Computer Science undergraduate at JIIT with a deep interest in backend systems, AI infrastructure, and applied ML research. I build production-grade systems that don't just work in demos — they scale, handle concurrency, and solve real problems.
My research spans Retrieval-Augmented Generation, diffusion models for motion synthesis, and graph neural networks for geospatial analysis — two of which have been published or submitted to IEEE/SCOPUS conferences.
Outside research, I contribute to open source through GSSoC, enjoy competitive programming on LeetCode (1720+ rating, top 5%), and mentor peers in DSA. I'm looking for roles where engineering rigor and intellectual curiosity are both valued.
Quick Facts
Where I've Worked
Deloitte
Technology Consulting Simulation
Virtual Experience • Forage
- ▸Completed technology consulting simulation focusing on tech dashboards and enterprise data analysis.
- ▸Developed analytical frameworks for enterprise-level data solutions.
GirlScript Summer of Code
Open Source Contributor
Open Source • Remote
- ▸Engineered security enhancements for Flutter authentication via OTP expiry and Twilio API integration, patching critical vulnerabilities and reducing unauthorized access by 40%.
- ▸Redesigned ShopSmart profile UI using Material Design and Lottie animations, boosting user engagement by 25% and increasing average session duration by 15%.
- ▸Merged 15+ pull requests across 10+ repositories, resolving performance bottlenecks and security vulnerabilities; changes reviewed and approved by project maintainers.
JPMorgan Chase
Software Engineering Simulation
Virtual Experience • Forage
- ▸Completed advanced software engineering simulation covering Kafka, Spring Boot, and REST API development.
- ▸Built real-time data streaming pipelines and implemented financial dashboard visualizations.
Coforge
Software Engineering Intern
Internship • Hybrid / On-site
Details to be updated soon.
Things I've Built
From research implementations to production systems — each project reflects a commitment to engineering depth.
ContextCore
RAG Document Intelligence Platform
Production-grade RAG system with hybrid retrieval, Redis caching, and real-time streaming — reducing query latency from 2.5s to 4ms.
- ▸Hybrid FAISS + BM25 retrieval fused via Reciprocal Rank Fusion
- ▸Redis TTL caching: 2.5s → 4ms query latency
- ▸Per-user isolated FAISS vector stores with JWT auth
- ▸SSE token streaming for sub-second real-time responses
RA-MDM
Rhythm-Aware Motion Diffusion Model
Research implementation of a rhythm-conditioned diffusion framework achieving 0.94 beat alignment — 6.5% improvement over HMDM baseline.
- ▸0.94 beat alignment across 4 dance styles
- ▸18% lower FID vs RNN/VAE/GAN/HMDM baselines
- ▸Style-adaptive FiLM layers + learned noise schedule
- ▸Published on ResearchGate with ablation studies
GAT Land Cover
Graph Attention Network Classification
End-to-end GAT pipeline on EuroSAT multispectral dataset achieving 76.9% accuracy and 100% F1 on water body detection.
- ▸SLIC superpixel graphs with 21-dimensional node features
- ▸100% F1 on water body detection via NDVI thresholding
- ▸2-layer GAT with 8 attention heads
- ▸Submitted to IC3-2026 (IEEE SCOPUS-indexed)
MindLink
AI-Powered Study & Wellness App
Flutter app combining AI chat, smart study summaries, focus mode, and mental wellness tracking with secure real-time messaging.
- ▸AI-powered study summaries using Gemini API
- ▸Real-time secure chat with Firebase
- ▸Focus mode + mental wellness tracking
- ▸Clean MVVM architecture with modular components
Gold Price Predictor
Investment Advisor with Soft Computing
Intelligent gold price prediction system using soft computing techniques to generate actionable investment signals.
- ▸Soft computing techniques for price forecasting
- ▸Buy/sell investment signal generation
- ▸Interactive Streamlit dashboard
- ▸Historical pattern analysis
QuizForge
Gamified Flutter Quiz App
Dynamic quiz app with real-time scoring, adaptive difficulty, dark-themed interface, and modular architecture for iOS and Android.
- ▸Adaptive difficulty scaling
- ▸Real-time scoring with leaderboard
- ▸Dynamic categories via Open Trivia API
- ▸Cross-platform iOS and Android
Technical Stack
Technologies I use to build, research, and ship.
Languages
Backend & Systems
AI / ML Research
Mobile Development
CS Fundamentals
Tools & Dev
Milestones
LeetCode Elite
Rating · Top 5% Global
Solved 1000+ problems across LeetCode, GeeksForGeeks, and Coding Ninjas. Consistent performance in contests.
Published Research
Papers · IEEE + ResearchGate
RA-MDM published on ResearchGate. GAT Land Cover Classification submitted to IC3-2026 (IEEE, SCOPUS-indexed).
GSSoC 2025
PRs merged · 10+ Repos
Open source contributions reducing unauthorized access by 40%, improving UI engagement by 25%. All PRs code-reviewed and approved.
Industry Simulations
JPMorgan Chase + Deloitte
Completed advanced engineering simulations covering Kafka, Spring Boot, REST APIs, and enterprise tech dashboards.
Community Leadership
Programming Hub · JIIT
Organized campus-wide coding contests as Volunteer at Knuth Programming Hub, mentoring peers in DSA and competitive programming.
Problem Solving
Problems across platforms
Consistent daily practice. Strong foundation in dynamic programming, graph algorithms, trees, and system design.
GitHub Activity
Contribution Activity
Pinned Repositories
ContextCore
RAG-powered document Q&A with hybrid FAISS + BM25 retrieval, Redis caching, FastAPI backend.
Rhythm-Aware-Motion-Diffusion-Model
Official implementation of RA-MDM for rhythm-synchronized dance sequence generation.
GAT-Land-Cover-Classification
Graph Attention Network for land cover classification on EuroSAT multispectral dataset.
MindLink
AI-powered secure chat app boosting study productivity with smart summaries and wellness tracking.
Let's Build Something
I'm actively looking for SDE internships and full-time roles in backend, AI/ML, and full-stack. Open to research collaborations too.
Reach out — I respond fast.
Designed & built by Aditya Chaturvedi · Next.js · TypeScript · Tailwind · Framer Motion
© 2026 · JIIT Noida