Hi, I'm Shivansh Rana, a Software Engineer passionate about building intelligent systems that combine AI, backend engineering, and scalable software.
🎓 B.Tech Computer Science @ Delhi Technological University (DTU)
💼 Software Development Intern @ Hero MotoCorp (Jun – Jul 2026)
🚀 Incoming Software Development Intern @ Uber (Aug – Dec 2026)
I enjoy building products at the intersection of LLMs, Agentic AI, Retrieval Systems, Distributed Systems, Backend Engineering, System Design, and Full-Stack Development. Whether it's designing scalable architectures, developing AI-powered applications, or exploring emerging technologies, I'm always looking for problems that challenge me to learn something new.
Beyond engineering, competitive programming has strengthened my problem-solving mindset, while music has taught me creativity, discipline, and confidence. Together, they've shaped the way I approach every challenge - with curiosity, consistency, and a willingness to step outside my comfort zone.
| 🏅 NTSE Scholar |
|
|
| 💻 1000+ Problems Solved on Leetcode | 🏆 Flipkart GRiD National Finalist | 🥈 Adobe India Hackathon Finalist |
| 🥈 Meta Hacker Cup Round 2 268 Rank | 🌍 150+ CSES Problem Solved | 🏆 Amazon ML Summer School 2026 Selected |
| 🎤 Top 12 – The Voice Kids | 🎙️ Top 100 – Indian Idol |
"Technologies and domains I'm currently investing the most time in"
 LLMs • Agentic AI RAG • Agentic Rag Prompt Engineering MCP • AI Agents |
Spring Boot Distributed Systems REST APIs Microservices |
Azure • GCP Docker • Devops Kubernetes CI/CD Pipelines |
AI Products Developer Tools Full Stack Apps Scalable Systems |
"The right tool depends on the problem"
"Building products that combine AI, scalable software engineering, and great user experiences"
|
A production-grade retrieval framework that uses AI agents to dynamically orchestrate the entire retrieval pipeline—from query understanding and chunking strategy selection to retrieval, reranking, and response generation. Unlike traditional RAG pipelines, RARE intelligently chooses the optimal retrieval strategy at runtime, making retrieval adaptive, modular, and evaluation-driven.
|
An intelligent web platform that combines Computer Vision, Search Algorithms, and Interactive 3D Visualization to recognize, solve, and teach Rubik's Cubes. Built for both beginners and speedcubers, the platform transforms a physical cube into an interactive digital experience with real-time recognition, optimal solving, and guided visualization.
|
|
Building something exciting...
Always experimenting. Always shipping. |
"Consistency beats intensity."
|
|
|
|
|
|
