Open to opportunities

Prithvi Reddy

@prithvireddy

Message

I build full-stack, real-time, and AI-powered applications from end to end.

India

Message

What I'm looking for

I'm looking to build practical full-stack, backend, or AI products where I can apply my work in real-time systems, RAG, machine learning, and scalable application development.

I built a Solana token swap platform with Next.js, TypeScript, PostgreSQL, Prisma, and Jupiter Aggregator APIs, supporting live quotes and on-chain swaps for SOL, USDC, and USDT.

I also engineered GoDraw, a real-time collaborative whiteboard using WebSockets, Node.js, JWT-authenticated connections, and PostgreSQL-backed room and canvas state. I structured it as a Turborepo/pnpm monorepo with four applications and six shared packages.

For an M&A research agent, I indexed 11,500+ chunks from 389 publications in Pinecone and built hybrid RAG retrieval with LangChain, Neo4j, Sentence-Transformers, parent-child chunking, and Cypher query generation.

During my AICTE Virtual internship, I developed a PyTorch CNN forest-fire detection pipeline that achieved 90+% held-out test accuracy across 5,000 labelled satellite images. I enjoy applying strong data structures, system design, backend, and machine learning foundations to practical products.

Experience

Work history, roles, and key accomplishments

Artificial Intelligence and Data Analytics Intern

AICTE

Apr 2025 - Present (1 year 4 months)

Developed and evaluated a CNN model in PyTorch to detect forest fires from satellite imagery, achieving 90+% classification accuracy. Built an end-to-end image classification pipeline including preprocessing, augmentation, training, and evaluation over 5000 labelled images.

PyTorch CNN data preprocessing Model Evaluation