Prapti Kodachi

Prapti Kodachi

AI Engineer in the Making | Generative AI, Agentic Systems, RAG & Backend Engineering

Kalaburagi, Karnataka, India


About

I'm a Computer Science undergraduate who loves turning ambitious ideas into working products. My interests sit at the intersection of AI engineering, Generative AI, agentic systems, backend infrastructure, and full-stack development. I enjoy going beyond building features — understanding how systems work under the hood, experimenting with emerging technologies, and creating solutions that are useful, scalable, and thoughtfully designed. From building a vector database from scratch in C++ and exploring retrieval-augmented generation to developing full-stack applications and AI-powered agents, I’m driven by curiosity and a bias toward execution. I’m especially interested in how intelligent systems can move beyond answering questions to reasoning, using tools, learning from context, and taking meaningful action. I believe growth comes from building things, asking better questions, and stepping into rooms where there is more to learn. I’m always looking to challenge myself, collaborate with ambitious people, and turn what I learn into something real. Currently exploring: LLM engineering, advanced RAG, AI agents, system design, and production-ready software architecture. Open to opportunities in AI Engineering, Generative AI, Backend Engineering, and Software Development.

Skills
Python
Java
JavaScript
React
Next.js
Firebase
TensorFlow
LangChain
PostgreSQL
MongoDB
CI/CD
Node.js
RAG Pipelines & Evaluation
Vector Search (HNSW
Qdrant
Ollama (Llama 3.2) LLM
Gemini API
Prompt Engineering
ML Fundamentals
Agentic AI (n8n)
NLP
Semantic Retrieval
Projects

VectorDB — A Vector Database Engine from Scratch

Designed and implemented a vector database from scratch in C++, exploring efficient similarity search and the fundamentals of AI retrieval infrastructure. Implemented HNSW, KD-Tree, and brute-force search approaches, with support for cosine, Euclidean, and Manhattan distance metrics. Integrated a web-based interface and local LLM-powered Retrieval-Augmented Generation (RAG) using Ollama to demonstrate how vector search enables semantic retrieval and context-aware AI responses.

C++17
HNSW
KD-Tree
Similarity Search
Vector Embeddings
RAG
Ollama

MediMeet — Panchakarma Healthcare Booking Platform

Built a full-stack healthcare platform connecting patients, therapists, and administrators through dedicated dashboards and streamlined appointment workflows. Implemented specialty-based booking, pre-consultation video calls, session tracking, payment workflows, and document verification. Designed a responsive, modern interface and integrated authentication, database management, and communication services to support end-to-end healthcare service delivery.

Next.js 15
React
TypeScript
Tailwind CSS
shadcn/ui
Clerk
PostgreSQL (NeonDB)
Prisma ORM
Vonage
Vercel
Education

Bachelors in Technology

Sharnbasava University Kalaburagi · 8.74

Sept 2023 to Aug 2027